OpenOffice Bloated?

cygnusx writes "ZDNet's George Ou has been writing a series of posts about Open Office bloat. Includes some interesting system usage comparisons" From the article: "Even when dealing with what is essentially the same data, OpenOffice Calc uses up 211 MBs of private unsharable memory while Excel uses up 34 MBs of private unsharable memory. The fact that OpenOffice.org Calc takes about 100 times the CPU time explains the kind of drastic results we were getting where Excel could open a file in 2 seconds while Calc would take almost 3 minutes. Most of that massive speed difference is due to XML being very processor intensive, but Microsoft still handles its own XML files about 7 times faster than OpenOffice.org handles OpenDocument ODS format and uses far less memory than OpenOffice.org."

Comments

"Essentially" the same data?

[LeonGeeste]

I seriously doubt that. I was working with files similar in size to the ones discussed in the article just last night, and I got completely opposite results. OO.org took half the time to load that Excel did and took up just over half the space for the files. I really don't know where they get these numbers. Probably a biased test with fundamentally different data. I hate trying new software that does the same thing, and I am by no means tech-savvy, but even I can see that OO.org runs laps around any MS product for my uses. I swear, this must be someone shilling for Microsoft.

Re:"Essentially" the same data?

[BoxRec]

My personal experience is the opposite, Micro$oft Office loads and runs a lot quicker, however it also crashes a lot more often.

Re:"Essentially" the same data?

[mnmn]

I dont know how you got the opposite results.

I installed OO 2.0 on my machine to check the updates, and to see if its speed is up to snuff. Issues with compatibility are gone but it is more than twice as slow while opening files. (I'm not using quickstarters for OO or MSO).

Heck since I'm reporting these results, I MUST be a microsoft shill too I guess.

Re:"Essentially" the same data?

[Anonymous Coward]

My company uses OpenOffice on literally hundreds of PCs. I use OpenOffice exclusively, save for the rare time we in IT use MSFT Office to open up a document for a user. (No, OpenOffice is not perfect at converting but that's OK with us).
OpenOffice is free (as in beer) and easy to use. The drawback? It's so friggin SLOW, SLOW, SLOW.

Real world use, real world PCs, real world users. OpenOffice is painfully slow. I have tested OpenOffice on dozens of PCs, users and my own, PCs and notebooks. Slow is as slow does, and OpenOffice does it slowly.

Kill the QuickStart process THEN try to open OpenOffice. UUGH! SLOW!!!!

Re:"Essentially" the same data?

[Cereal Box]

Hmm, you didn't provide any hardware/software specs OR timing/memory data (so others could confirm your work), but your results are nonetheless "informative". It must be because your results were in OO's favor.

Re:"Essentially" the same data?

[Otter]

I swear, this must be someone shilling for Microsoft.

Well, at least that makes more sense than the guy modded above you who thinks this is someone shilling for Intel!

Anyway -- the test files are linked in his post. I don't have OO installed at work, but would someone with it and Office care to take a look at them, and move this discussion to facts?

Re:"Essentially" the same data?

[fishybell]

Well, don't believe it? Benchmark it yourself.

He provided the test data here and here

Re:"Essentially" the same data?

[plover]

Hmm. I've been running MS Office 2003 for over a year and have yet to experience a single crash with Word or Excel. I've had Outlook freeze up numerous times, but virtually all of those problems have their roots in our Exchange server (and the seriously mismanaged overload they've piled on it.)

Re:"Essentially" the same data?

[jsailor]

I'd have to second that. I regularly use both MS Office and OpenOffice Beta2 and do not notice significant differences between the two. Some of my spreadsheets are fairly complex and some of my documents are large, but most are the garden variety items that get passed around the Fortune 500. Both applications work fine and neither of them cripple my very modest laptop with 512 MB of RAM.

I thought finding obscure weaknesses in products and crafting tests to exploit was a task for vendors not industry commentators.

Re:"Essentially" the same data?

[BarryNorton]

Hmm. I've been running MS Office 2003 for over a year and have yet to experience a single crash with Word or Excel. I've had Outlook freeze up numerous times...

Likewise. What's more I've had (win32) xemacs, yap, (cygwin) xfig and ghostscript crash inumerable times this week, let alone this year...

Re:"Essentially" the same data?

[sinan]

Even then , would you rather spend $500 on Microsoft Office or extra memory and a CPU to get a better machine to speed up all the applications. Paying for one piece of software usually does not pay.

Re:"Essentially" the same data?

[madman101]

With the exception of Outlook, Office 2003 has never crashed on me, even when handling huge files. On the other hand, when we evaluated Open Office, we couldn't get it to stay up for more than 1/2 hour, and when it did work it was unacceptably slow.

Re:"Essentially" the same data?

[/ASCII]

The article author provides a download of the documents he used, and invites everybody to download it and try it out. You are referring to private data that you claim show the opposite. I know who I'm more inclined to belive.

But I'll admit to beeing too lazy to actually find out for myself.

Re:"Essentially" the same data?

[morgan_greywolf]

On what platform/hardware/etc., though?

I've found the OOo runs much quicker on my Linux box than it does on Windows running on the same box. As for how much memory it consumes, well, the thing about that is that while Calc uses much more memory than Excel, when you load 'Calc', despite appearances, you are in fact loading almost the ENTIRE office suite into memory, including the word processor, database front end and presentation graphics application. There is not much different architecurally between OOo and the older StarOffice 5.x -- they've just gotten better at hiding the fact that you need to load the entire office suite to load one application.

So a truer comparison would involve starting Word, Excel, PowerPoint and Access, watching how much this entire toolset takes up in memory, and then load the Excel and Calc files and see the difference.

Re:"Essentially" the same data?

[zagmar]

Is it possible that part of the issue has to do with MS Office using certain aspects of Win XP that OpenOffice does not have access to?

Re:"Essentially" the same data?

[gowen]

So a truer comparison would involve starting Word, Excel, PowerPoint and Access, watching how much this entire toolset takes up in memory, and then load the Excel and Calc files and see the difference.

Well... no.

Just because the design of OO.o is completely braindead, that's no reason to handicap the competition to make it look better. If Excel is smaller than Calc, say so. If Word is smaller than Writer, say so. If Word+Excel+Powerpoint combined are about the same as the OO equivalents combined, then say that, but most of the time people want just Word, or just Excel, or just Powerpoint.

Re:"Essentially" the same data?

[Weh]

I have an old P3-500, on that machine oo's spreadsheet app takes forever to render a simple chart, no such problem with xls. Other than that I use oo often enough (I don't have ms office installed anymore)

Re:"Essentially" the same data?

[Pentavirate]

I once tried to run Star Office back when it was free. It wouldn't even run on my, admittedly, slow machine, but I was running Office 2000 without a hitch. Isn't OO.org based off of this same Star Office code? Is anyone surprised?

Re:"Essentially" the same data?

[mikefe]

1. It is not fair to compare based on file size. Not only are OOo files compressed, but different data that is the same size uncompressed can have drastically different processing times. Think of the difference of one page full of vector graphics, tables and a little text compared with 3 or 4 pages of text.

2. It is a known problem that OOo takes a while to start. Staroffice (at the point when Sun bought it) was made by a German company. Most of the internal functions are named in german, and use abbreviations that are not obvious. The fact is that each version of OOo has been getting smaller and faster. OOo 2.0 is the same. If you run OOo 1.1.4 and OOo 2.0 side by side on windows, the 2.0 version uses about 10MB less memory when both have nothing open.

3. Since it uses more memory, it has a higher chance of being swapped out when you switch to another program for a while. A good way to see this in a short period of time is to run a torrent in the background (seeding or just downloading). Leave an OOo window open and use another program for 20 or more minutes. When you switch back to OOo it can take 10-40 seconds (depending mostly on the speed of your hard drive and amount of memory available) for the window to redraw.

If you are using OOo often enough to keep it in memory it is very snappy. But if it gets swapped out, then you will notice a speed degredation.

4. In my experience with small files (less than 200 records in a spreadsheet and 1 - 4 page documents) OOo takes longer to open and save files. I usually work with .csv, .xls, .doc, and of course .odt and .ods files.

Re:"Essentially" the same data?

[arkanes]

One common cause for this discrepency is that Windows does pre-caching and pre-binding for commonly used applications. When you first install Firefox or OO, it will be slower, but if you don't use IE or Office for 6 months, while you use the alternatives regularly, the Microsoft apps will be slower after a while. IE takes *forever* to load on my laptop on the rare (once or twice a year) occasions I fire it up.

Re:"Essentially" the same data?

[DaEMoN128]

Thanks for the links. I do have one question, I cannot test on my work computers, does the xml document for ms office perform as well in the oo.org? If we want a true and fair test, we should be using the same computer and file in the two different programs. Opendoc may take longer, but it can be done in OO.org where MS office cant do it. Does the test follow the same pattern of results when using the MS office file in openoffice? Is OO.org still slower? I can see this as possible, but I would like to see a proper test where only one variable is tested instead of two. Untill that happens... this test is just spouting off. It may or maynot be true because we dont know how openoffice does against ms office in .doc files. I will testify that it takes longer to load on my Vaio laptop. But that is not a big deal to me. I havent checked the memory usage yet. OH, yeah, what verison of office and oo.org was he using?

Re:"Essentially" the same data?

[Bluejay42]

Is it time yet to admit that Excel is an excellent piece of software? It has managed to stay true to its core competencies (calculations) while gaining many new audiences. I have used it in the past year for stock analysis, dynamic web queries (a simple Crystal Reports), and site wireframing. While many look at the insecurity of macros, they are enormously useful for the financial community and advanced data analysis.

The great thing about open source though, I bet there is someone *right now* using the test files provided by this author to improve the Open Office parsing routines. Gotta love it.

Re:"Essentially" the same data?

[Anonymous Coward]

you're using MSO quickstarter since you've installed.

Re:"Essentially" the same data?

[MikeMacK]

Hmmm...on my Linux box OpenOffice really outperforms Excel...oh wait...

Re:"Essentially" the same data?

[Anonymous Coward]

but even I can see that OO.org runs laps around any MS product for my uses

You've got to be kidding me. OO has never been faster than any version of MS Office I have ever tried. Without that "booster" application sitting in your system tray the individual applications take usually about 2x as long to load and be in a usable state as the equivalent MS Office application.

Now, on to some real numbers. I'm timing this with my watch so you'll have to forgive the ~1 second resolution. I perform each test several times to ensure that disk I/O doesn't taint the numbers.

A random excel spreadsheet on my desktop that calculates some manufacturing costs.

XLS format - 1980 kb

Opening in Excel XP (this includes the time to load excel)
Less than 1 second

Opening XLS file in OO (Calc already loaded)
5 seconds

Ok. To be fair we should save the spreadsheet in OO format so the converter isn't required. Let's test save times first though.

Save as new XLS file under Excel XP
3 seconds

Save as new XLS file in OO
3 seconds

Save as new native file in OO
4 seconds

Ok. Let's see how long it takes OO to open a document saved in native format.

Open ODS file in OO (Calc already loaded)
7 seconds (not surprised - XML processing is sssllllooooowwwww)

And finally, on to memory usage with said spreadsheet loaded:
OO Calc - 67 meg
Excel - 15 meg

I won't even mention the issues with things like the noticable delay between the time you click the menu and the time it appears. Don't get me wrong - OO 2.0 is a nice office suite but don't claim it "run laps around" MS Office. That isn't true by any stretch of the imagination.

Re:"Essentially" the same data?

[starwed]

That's probably why OpenOffice installs the quickstart by default. (I turn it off because I almost never deal with "office documents" at home. But considering the crap that windows loads by default, that doesn't exactly make it an unfair comparison. ^_^)

Re:"Essentially" the same data?

[jambarama]

Yeah OO.org is slower, but you really are using a quickstarter for excel. A lot of office is loaded on startup. So while this doesn't completely account for OOo slowness, it does a fair amount. Likewise Koffice is faster in KDE than OO.org

This is the same with IE. IE boots faster than either firefox or opera, on a coldboot (if the programs haven't been run before). It doesn't mean IE is worth using, just boots faster.

Re:"Essentially" the same data?

[coronaride]

I'm not sure about the specifics of pre-caching, but there is some truth in your comment. However, your comment is still what I would consider a troll or a nag. At the very least, it's a good example of blind ignorance from a Linux apologist. If using a prefetch causes my documents to load faster, then that would seem to me to be good design. Your argument is analogous to the folowing:

Your car is faster than my bicycle, but only because it uses a complicated internal combustion engine. If you didn't use your car for a couple of years, it wouldn't start up very fast, would it? I'd be halfway down the block before you even got your car in gear...

Re:"Essentially" the same data?

[CyricZ]

Indeed, those products are rather lousy on Windows. But then again, Windows is vastly different from UNIX, and the UNIX layers on Windows (like Cygwin) aren't very good.

On Linux, BSD or Solaris, for instance, such programs work very well. Indeed, one only needs to look at Kylix to see the reverse being true. The port of the Delphi IDE, using Wine, was terrible. It crashed terribly often, not that it was completely stable under Windows, mind you.

Re:"Essentially" the same data?

[toleraen]

Windows does pre-caching and pre-binding for commonly used applications

After installing OO, check your system tray. It does the same thing as Office does by providing a "Quickstarter" application at startup.

Re:"Essentially" the same data?

[jahudabudy]

If using a prefetch causes my documents to load faster, then that would seem to me to be good design.

I think you misunderstand the OP (although, its possible I do). My take on his comment is that whichever application you use most commonly is going to be improved thru pre-caching and binding. So, if the testers use Excel often, their Excel load time is going to be faster than my load time if I never use Excel, even on otherwise identical machines. His point, I believe, is that if the testers did not account for this, then the Operating System's behavior of pre-caching and binding could play a significant enough factor in the load speeds to completely overshadow any differences in the Applications' respective abilities.

Re:"Essentially" the same data?

[Anonymous Coward]

no it's more to do with OO sucking, and MS Office being (apart from a few niggles) pretty darn slick.

Re:"Essentially" the same data?

[Anonymous Coward]

re: swapping and torrent

Just dont use azureus!

Re:"Essentially" the same data?

[arkanes]

This is exactly my point, and the previous responded did in fact misunderstand or misrepresent my point. It has nothing to do with Linux apology, and has nothing to do with the relative performance of the application. It's an explanation (there are other contributing factors) for the different experiences of users.

Re:"Essentially" the same data?

[utnow]

So... I should load Macromedia Fireworks, Dreamweaver, Flex, Coldfusion, Studio, JRun, Freehand, Authorware, and Contribute every time I want to edit a flash file with MM Flash?

Or should Photoshop also load up Illustrator, GoLive, and Premier?

Naw... there's no bloat in OO.org

Re:"Essentially" the same data?

[eno2001]

Some of this is also arrogant assumptions on all sides. The pro-MS camp is going to test Office under Windows against OO.o under Windows. However, OO.o is a port. It wasn't written for Windows, so there are bound to be some performance issues. Likewise, the pro-OO.o camp will post their various figures proving otherwise while running OO.o under Linux or some other *nix. All the while not taking into account that "Joe Average" at home isn't going to be running the optimal OS for OO.o. I've even seen some weird responses on the ZDNet site. Like the guy who compared OO.o running under Linux with MS Office running under Wine and providing memory usages stats from that. That's just plain silly. Wine has a LOT of overhead, so it's unnfair comparison in much the same way that comparing MS Office on Windows to OO.o on Windows is. The real truth would be found in comparing MS Office on Windows with OO.o on a *nix and then providing the caveat that each Office suite is made to run well on a particular platform. That's the most realistic view. Of course when the AJAX based Office suites pop up all over the place, this will all be moot.

Re:"Essentially" the same data?

[Martin Blank]

Benchmarks for me, performed on:

Dell Latitude C840
1.4GHz CPU (plugged in, so full speed)
1GB RAM
5400 RPM 60GB hard drive
Windows XP SP2

Results:

Microsoft Excel
Application opening time: ~3 seconds (no preloader used)
Application opening RAM: 11MB
File opening time: 80 seconds
RAM usage with file open: 45MB

OpenOffice.org 2.0 Calc
Application opening time: ~4 seconds (no preloader used)
Application opening RAM: 35MB
File opening time: 379 seconds
RAM usage with file open: 240MB

Factors:
Application opening time: Effectively a tie
Application opening RAM: Calc uses about 3x more
File opening time: Calc takes about 4.75x as long
RAM usage with file open: Calc uses about 5.33x more RAM

Of course, this probably won't affect that many people. I'm a bit of a spreadsheet fiend, and even my largest work is rarely more than a few hundred KB. I did find it interesting that when trying to open the Excel XML file, Calc could not complete the opening, showing a General I/O failure.

Re:"Essentially" the same data?

[shark72]

"It is a known problem that OOo takes a while to start. Staroffice (at the point when Sun bought it) was made by a German company. Most of the internal functions are named in german, and use abbreviations that are not obvious."

While I understand that you were not trying to be an open source apologist, this statement is the epitomy of the frustration that many mainstream users have with open source:

Office worker: "why does Open Office take so long to load?"

IT guy: "That's because the routines were written by a German guy in his free time. I'm sorry, little-miss-everybody-should-speak-English, but this poor guy was working for free. What do you expect?"

Office worker: "what does the German language have to do with this?"

IT guy: "Your PC was built in Austin, Texas. German is its second language. See this routine here, öffnenSiediegroßeAkte()? Your American PC doesn't know what that means, and has to consult a dictionary each time it sees it. There's a group of teenagers translating it into English. They work on one word each for greater safety. One of them saw two words of the program and spent several weeks in the hospital."

Office worker: "So, what do I do about it?"

IT guy: "Have a little more tolerance for the global community in which we live, and worship the holy light of open source that's shining out of Richard Stallman's ass. Oh, and consider dying your hair blond, adopting a schnauzer, and carrying a riding crop. Open Office seems to like that."

Office worker: "Dork."

If this continues, I think it will inevetably lead to new ad campaigns like:

Microsoft Office: We won't coerce you into adopting a schnauzer!

Re:"Essentially" the same data?

[civilizedINTENSITY]

More like, one bicycle is locked up out front, one bicycle is in the back of the garage behind crates with piles of stuff on top, and yet a third bicycle is somewhere in the attic, we think. I yell "go", and time who gets around the block faster. Storage placement and access is not a fair test of the design of the bicycle.

Re:"Essentially" the same data?

[civilizedINTENSITY]

Could you run that same test for Gnumeric?

Re:"Essentially" the same data?

[Parham]

It's not just azureus, try downloading and then seeding any linux distro offered on torrent and watch what happens when you try to do anything with big software shortly later.

Re:"Essentially" the same data?

[Anonymous Coward]

While i see where you are coming from I have to disagree with you. It doesn't make any sense at all to compare MS Office under windows and OO.o under linux. If you want people using Windows to switch from Office to OO.o they will being doing so under Windows. Therefore the speeds/memory use under windows is what needs to be compared.

Re:"Essentially" the same data?

[plover]

I have no problem with Excel. It does exactly what I want it to do, and has for many years. About 10 years ago I developed a "quoting application" that did materials computations, profits, etc., and then with the click of a button would generate a printable version of the quote for the customer and would also fax it directly to him. Was some of it ugly? Sure. Was it tied to a weird fax interface, and completely non-portable? Yup. But it was possible then, and it still works today (except for the faxing part.)

Re:"Essentially" the same data?

[letxa2000]

My experience has been that OpenOffice LOADS (i.e., itself) much slower than Microsoft Office--I'd dread having to open OpenOffice Writer and would usually wait for a few tasks to pile up so that I could make the most of the fact that I was opening up Writer. Microsoft Word would open up in just a couple seconds.

HOWEVER, once the application itself is running, I found that load times were not significantly different. Granted, I compared Writer to Word--I'm not a heavy spreadsheet user. I also find that the current version of OpenOffice (or whatever ships with Fedora Core 3) is faster in loading itself than whatever version of OpenOffice I was using before (1.0, perhaps?).

The comparison became irrelevant to me, though. I switched to OpenOffice last year when I was writing a book. The book was mostly text but did have its share of diagrams and graphics. Eventually, Word would just bomb out. I literally couldn't make progress on the book since Word would randomly crash out and I'd have to rewrite some amount... it'd happen several times a day and just plain annoying. So I tried OpenOffice. Took me about half a day to adjust for some formatting idiosyncracies that Writer had with the 200+ page Word document, but once those minor tweaks were done, I cruised on through writing the rest of the book. Writer SXW files are also much smaller than equivalent Word documents from what I've seen.

The book ended up with 348 pages... OpenOfffice Writer didn't crash even once for the remaining 10 months of writing and editing I did.

The speed comparison is irrelevant to me. The fact remains, I was able to author my book in OpenOffice Writer while Word constantly crashed once the document got to some filesize and/or some level of graphics.

Re:"Essentially" the same data?

[rdoger6424]

Are you using a beta or stable release of OpenOffice? I think that there are two different versions (Beta being the buggier version, stable being bug free, persumably)
And yes, OO.org is based off of star office.

Re:"Essentially" the same data?

[amliebsch]

I think he's referring more to Windows' trait of moving the data for the most commonly used programs to defragmented sectors on the outer edge of the hard disk platter. The quickstarter may pre-load parts into memory, but it doesn't improve disk performance.

Re:"Essentially" the same data?

[yoyhed]

I have used Office 2003 and 2002 extensively and never experienced a single crash (although now I use OpenOffice because I'm too lazy to put a CD in my drive to install Office after I format).

I'm sick of people who will make concessions about a Microsoft product being good, and then just throw in at the end "but its m$ so it crashes all teh time lolol!". Very original with the dollar sign in Microsoft's name, now we all know we can take you seriously.

Re:"Essentially" the same data?

[ciw42]

Nope, the previous poster is right.

The reason MS Office performs well, seems to load quickly and use less memory is the fact that it uses mostly libraries that are loaded and used by Windows from startup.

A true cross-platform application like OOo won't make use of such platform specific features for the sake of portability, and pretty much all of the bloat you are refering can be attributed to this. Yeah, there's going to be some legacy code knocking around, but that's also true of MS Office.

However, once you're run *any* of the OOo apps for the first time after a reboot, subsequent startup/reload time is actually around the same as MS Office.

If you'd rather spend £250+ on software (or run illegal copies) just to save a few seconds when you open your first document of the day, then go ahead. I don't hate or have a problem with MS Office, but I genuinely prefer OOo for day to day use. To my mind and way of working (and the many others I've introduced to OOo), it's UI is more streamlined and better organised, so even if the MS apps were free, I'd still go with OO.

Re:"Essentially" the same data?

[Anonymous Coward]

I can't speak of versions beyond MS Office '97, but I can transplant the install from Windows 98 to XP and run that from scratch and it opens in a few seconds. There's no way it's caching startup stuff then. Openoffice is a big bloated pigdog. I wish it wasn't so but there you go.

Re:"Essentially" the same data?

[Pentavirate]

I have used the "stable" release but not too much. My point is that it should be no surprise if OO.org is bloated because it's based off the old Star Office code which was some of the most bloated softare I'd ever used.

Re:"Essentially" the same data?

[Anonymous Coward]

mod parent up! Its my birthday for gods sake.

Re:"Essentially" the same data?

[Martin Blank]

Gnumeric for Windows 1.6.0 RC1
Application opening time: ~3 seconds (no preloader used)
Application opening RAM: 19MB
File opening time: 53 seconds
RAM usage with file open: 397MB

This is for the SXC file. It crashed trying to open the Excel XML, tying up almost 2GB in resources before doing so. Gnumeric was the fastest loading of all three of them, but used the most resources. It also had a minor error ("Invalid attribute 'vertical-align', unknown enum value 'automatic'") when opening but I don't immediately see anything wrong with the data. The DOS box from which it spawns (?!) also says that the version of Pango library I'm using is buggy, but since if it's installed, it was installed by Gnumeric, that's not my fault.

Re:"Essentially" the same data?

[RWerp]

"2. It is a known problem that OOo takes a while to start. Staroffice (at the point when Sun bought it) was made by a German company. Most of the internal functions are named in german, and use abbreviations that are not obvious."

Geez, now I get it. And since Windows is made by an American company, it doesn't know foreign languages and has to check up each function name in the dictionary -- which slows down the program considerably. Thanks for explaining this mystery.

Re:"Essentially" the same data?

[skotte]

Oh. My. God. The 7 opening comments on this article (excepting the fFirst post) are actually *compliments* fFor Microsoft.

Bizarre, man. Bizarre.

Re:"Essentially" the same data?

[Anonymous Coward]

Where the fuck do you work that an office worker has to buy their own copy of MS Office, or their own RAM?

Maybe you should find another job?

Re:"Essentially" the same data?

[Neil Blender]

Funny story:

(No Excel benchmarks since I don't have a similar machine running windows)

On a ~2.4 P4 with 2G RAM running RHE3:

In every following case the processor was being 95+% utilized by OO.

OO2 and 001.14 perform about the same. The file took 5 minutes to open and maxed at 350M ram.

So, I save the test file as Excel in OO2. And OMFG! 1.03G of RAM is sucked up! It frees some of it up once the file is saved but now it's using 661M RAM. The file took about 90 seconds to save.

Anyway, I quit 002, relaunch it and use it to open the Excel file I just saved. It takes 45 seconds and uses 140M RAM. So I save this file as another Excel file and it takes about 15 seconds to save and the RAM only jumps 474M (and stays there even with only one 50M file open doing nothing.)

The test data file is like 3.7 Megs and the saved Excel file is 50 Megs.

LOL, it's funny because it's so stupid.

Luckily most of the data I get from people is in excel or tab delimited format. This 13 column, 16000 line file is nothing. I get 50 column, 50,000 line files all the time. If they were in opendocument format it would take forever.

PS Anyone know a mime-type I can use to get to OO plugin for firefox to open tab-delimeted text files as a spreadsheet?

Re:"Essentially" the same data?

[generic-man]

OpenOffice.org is based on StarOffice, which was available for Windows natively. The problem is similar to Mozilla: the application is not "native" to any one platform because it uses its own UI widgets, forms, and back-end data structure. It is no more a native Linux application than it is a "port."

Re:"Essentially" the same data?

[Burz]

I think it is becoming apparent that Excel uses a file-loading scheme that does not scan the whole file, nor load many portions of the file into memory until they are used.

With Calc and Gnumeric, there seems to be an assumption that the whole spreadsheet needs to be loaded into RAM before using it.

OTOH, I wonder how Corel Quattro would perform. It's another spreadsheet with a pedigree that goes back to the memory-constrained era, and competed directly with Excel on the basis of performance. I have a Linux version around here somewhere...

Re:"Essentially" the same data?

[Maxo-Texas]

I think George Ou would need to use both applications for the first time on a system which they had not been used before. Or George would need to use Openoffice and Word daily for a month. Since he uses microsoft products, his will already be set up to cache. Since he does not use Openoffice, it will not be cached.

If neither were used or both were used for a reasonable period before the test, then they might perform differently.

Knowing Microsoft, there is also the distinct possibility that it pre-caches it's own applications even if they are not used. And finally, there may be parts of their functionality (like text window management) built into the OS which cannot be used by system agnostic software. They would be -risky- to use, since Microsoft has changed such things in the past without warning breaking competing apps that used microsoft operating system features while not breaking their own applications.

All that being said, I speculate that OO programmers are more "theory" type and program pretty code that performs poorly while Microsoft programmers are more pragamatic programmers who are comfortable writing down and dirty and fast/efficient code in the 1% of the program where it's needed. Likewise, I bet Microsoft has better profilers. And finally, Microsoft has had a longer period to clean out grossly inefficient things.

OO will continue getting better. OO will continue to be free. OO will not lock you into an operating system. You will never have to "rent" OO. You will never be ratted out by OO to microsoft (Word hasn't done this but XP sends a lot of messages home and even reboots my system without asking me to apply patches). Finally, hardware and ram continue to get more powerful. In a couple years, the difference will be between 100 and 200 megabytes of your 8 gigabytes of memory and .2 and 2 seconds on your 8ghz (or dual core 4ghz) machine.

Re:"Essentially" the same data?

[SanityInAnarchy]

Where the fuck do you work that an office worker has to buy their own copy of MS Office, or their own RAM?

College.

Re:"Essentially" the same data?

[budgenator]

My wife took a college class on Office 2003, and it was obvious that the authors not only wrote the book in Word so they get points for eating their own dog food, but had split each chapter into a doc; so it appears that even the experts have trouble getting around this. Another thing I notice about windozers is they most tend to use one app at a time and shut their computers off; this makes them more sensitive to load times and boot times. When I fire my Linux machine up, I don't care about boot times because when it's up, it stays up for months. The same with applications, I frequently have four or five applications running for days or weeks.

Re:"Essentially" the same data?

[mikefe]

Geez, now I get it. And since Windows is made by an American company, it doesn't know foreign languages and has to check up each function name in the dictionary -- which slows down the program considerably. Thanks for explaining this mystery.

Ug, so it looks like you didn't make the mental connection about the german...

First of all, StarOffice was known for its bloat long before being bought by Sun. Before OOo 1.0 was released, a lot was done to clean things up, like the load of **** (flush twice) Star Desktop.

With the code already having a lot of bloat, with comments, function names (think german acronyms), in german doesn't help with understanding what everything does. This is slowly changing though.

So, use OOo because it is free and Free. Because of the functionality. Because the copy of MS Office on your friend's computer probably wasn't paid for. But not because it is fast. That will come with time.

Re:"Essentially" the same data?

[mikefe]

Yep, whether your bittorrent client is written in Java, Python or C++, if the amount of memory isn't at least big enough to fit the size of the files being seeded + memory for apps, then your apps will be swapped out in time.

Re:"Essentially" the same data?

[Kadin2048]

I agree, with some qualifications.

Excel is a pretty solid program. What I don't like is how some people use it. Because Excel has some database-like features built in, it leads people to attempt to use it as if it were (and in place of) a real desktop database, leading to really ugly, bloated spreadsheets and finicky cross-sheet references that break at the drop of a hat and are a major pain to track down.

I work regularly with a gigantic spreadsheet like this which is used as an internal financial planning and forecasting tool. The reason I've heard for originally using Excel is that they wanted something that would make graphs. Beginning from that premise, they piled all the data into a spreadsheet, and added sheet after sheet of subtotals forms, reports by week and month, reports by person, etc. It's truly hideous. In order to add a new person, you have to (by hand) modify each of the sheets, update the subtotal lines, etc., while in a real database program this would be relatively simple, if it even required any additional effort at all.

Excel is a great spreadsheet program for doing spreadsheets. What it's NOT is a desktop database, and far too many people are laboring under the impression that it is.

If Microsoft would quit loading all the query and PivotTable type features into Excel and concentrate on being a better spreadsheet than it is already, I wouldn't have any complaints. Even so, I suppose I grudgingly have to say that Excel is probably the least-offensive MS program out there.

Re:"Essentially" the same data?

[shark72]

"So, use OOo because it is free and Free."

This actually makes sense. Free as in beer... and who knows beer better than the Germans?

This fills me with a great sense of ichhasseschwierigestückescheiße... no real direct English translation, but Germans use it to describe those moments when one is angstful about open-source software. Remarkably compact language, that German is.

Re:"Essentially" the same data?

[ozmanjusri]

Bizarre, man. Bizarre.

Nah, it's pretty normal. Take a look at any recent story with a MS or OSS focus, and as soon as someone makes a post critical of Microsoft products, there'll be half a dozen;

"I've been running MS [product name] for over a year and have yet to experience a single crash with MS [product name]."

or similar type of posts. You often get modded down pretty heavily for the criticism too.
It looks like MS marketing keeps an eye on Slashdot.

Re:"Essentially" the same data?

[Anonymous Coward]

regardless, i had been using OOo for a few months, but recently switched back to ms office for 2 reasons.
a. excel loads much faster, and word loads slightly faster.
b. the thesaurus is better in word

excel was still better even after months of open office.

Re:"Essentially" the same data?

[lunatik17]

There's an easy way to test this. Open up regedit, search for a key named "EnablePrefetcher" and modify its value to 0. Go to C:\windows\prefetch and delete the files in there. Reboot, and the caching feature will be disabled.

Re:"Essentially" the same data?

[Quantam]

My experience is much like the other respondants' with respect to crashes. While it always hurts to lose a page or more of a term paper, Word and Excel (Office XP edition) crash only two or three times a year, for me (combined). Heck, I used to lose 1.44" floppies more often than that :P

Also, I've never used OpenOffice, so I can't compare it to MS Office, but I've always found MS Office to be very compact (i.e. Word only takes 16 megs RAM with my completed finance term paper open).

Re:"Essentially" the same data?

[Martin Blank]

I think it is becoming apparent that Excel uses a file-loading scheme that does not scan the whole file, nor load many portions of the file into memory until they are used.

Or it may just be that it's that much more efficient. I didn't detect any performance differences when moving around through the ~14 tabs and ~16K rows in either Calc or Excel. I didn't check that in Gnumeric, though.

Re:"Essentially" the same data?

[Achromatic1978]

I disagree with this. From the way I read it, he was complaining about the time taken to load a /file/ within the application, and that the delays were due to things like XML parsing, and so on, which no preload is going to prevent.

Re:"Essentially" the same data?

[0b11111010101]

I used to use Outlook 2003 all the time and it crashes on a regular basis. It also takes a LONG time to load once you get it full of emails, and it does not handle IMAP correctly. (No Exchange Server.)

Re:"Essentially" the same data?

[Achromatic1978]

With the code already having a lot of bloat, with comments, function names (think german acronyms), in german doesn't help with understanding what everything does. This is slowly changing though.

There are no polyglots on the development team?

Or is this 'even the developers don't know half the shit this code is doing!'?

Re:"Essentially" the same data?

[RichardX]

Sigged :P

Re:"Essentially" the same data?

[gordgekko]

Golly, could also be that many people never have a problem with Microsoft products. Like me. And if I'm being paid by Microsoft for this positive comment, I hope the cheque comes soon. My VISA bill is due.

Re:"Essentially" the same data?

[markdavis]

I tested with Linux & Athlon 4200+ & 1 GB RAM. I can't compare to MS-Excel because I don't use MS-Windows:

OO 2.0:

load sxc: 126 sec
save ods: 91 sec
open ods: 106 sec

So there is definitely some validity to OO being slow when opening very, very large spreadsheets. Of course, I don't know how important that really is, since most people will rarely, if ever, try to load/save spreadsheets with over 3 million cells. I use spreadsheets all the time and even the largest one I could find only takes 4 seconds to load.

As a comparison, Gnumeric took 11 seconds to open an xls version of the same multi-million cell spreadsheet. It is apparent that work does need to be done to further optimize OpenOffice in this regards.

Re:"Essentially" the same data?

[synthespian]

Excel had many problems regarding accuracy. That's why you don't see experts in statistics using it. Right now, it looks as though the professional statistics community has flocked to R, which is open source BTW (this does not regard other groups of heavy users of statistics, albeit in a limited way, such as medical doctors).
There are papers about this, google and ye shall find.

Re:"Essentially" the same data?

[Perl-Pusher]

I agree, I think that Excell is by far the best program Microsoft has ever produced. Unfortunately, the rest of office peaked in Office 97 in my opinion. I really like OpenOffice.org 2.0, but OOCalc does need serious attention. Gnumeric is my spreadsheet of choice on linux. OpenOffice is slow booting, but after that seems fine. I use it daily at work, nobody knows I am not using word or powerpoint.

Re:"Essentially" the same data?

[Tux007]

"Another thing I notice about windozers is they most tend to use one app at a time" True. After years off using Linux I'm still surprised, that I can have so many appliciations open on the 4 virtual desktops, without any problems with speed and stability. Under Windows I'm still used to close down applications all the time, because otherwise I'll get into trouble. If I can keep OO, my pim (I use Kontact), a console with mpg123 and a browser with a lot of tabbed windows open on different virtual desktops, I'm not complaining about speed. But it's true, if you click on a spreadsheet, while OO is not started, it takes some time.

Re:"Essentially" the same data?

[advb89]

Oh... that's because you forgot to hit the button labeled "ON" located on your Windows box. But seriously, i use both word and excel all the time on my high school student pc that our school provides, and have had it crash too many times to count. Does that mean i think it is bad software?? No, but with as many restrictions as the county has on my laptop, i don't blame it for freezing... (i can't even open Task Manager...)

Re:"Essentially" the same data?

[MadChicken]

Oh... my... goodness. That was absolutely fabulous.

Re:"Essentially" the same data?

[lkeagle]

So in other words, we don't know, and we can't make comparisons.

Until it's determined exactly what procedures MS goes through when it loads files, we can't make any comparisons.
Of course, we could go through OO's code and see exactly what it does.

All I can tell from reading just about ever post so far is that the following possibilities seem likely:
Windows likely pre-loads a portion of the Excel dll's at startup.
Excel likely loads just the portion of the file it needs to display at any given time.
Excel's format likely contains a large amount of meta-data so that it isn't required to ever scan the entire file.
OO Calc likely does not attempt to load pieces of its binary into memory until it is executed.
OO Calc likely attempts to load the entire data set into accessible memory upon startup.

So all we can conclude from ANY of this is that there was likely a completely different design that went into OO vs Excel. It seems that the OO developers felt that it was more important to have the data available in memory than to leave it on the disk and view a small subset of the sheet. Apples and Oranges...

In that light, what would a fair competition between the two applications entail? I would imagine that testing it's ability to sort or sum all the cells in the sheet would be an interesting test. Presumably, if the assertions I've listed above are true, then Excel will have to spend its time accessing the disk eventually. Of course, even this test can be skewed if Excel uses some kind of hardware acceleration to process the data that OO doesn't use because whoever compiled the binary didn't use that particular optimization. In any case, I would be just as skeptical of this test than I am of the first. It's incomplete and there is no control to compare to.

Re:"Essentially" the same data?

[Kelson]

Not sure if it's relevant, but as I recall Excel was developed outside MS and purchased to become a companion to Word. (This would explain why it's not called Microsoft Spread, as most apps that originated within Microsoft tend to use purely descriptive names like Windows, Word, Internet Explorer, Flight Simulator, etc. I never did figure out Bob, though.)

If correct, this could provide a degree of comfort to those who wish to preserve their MS-hating credentials and still admit that Excel is a damn good office app.

Re:"Essentially" the same data?

[dcam]

Love the Monty Python reference.

For thoe who don't recognise it:
http://www.pion.ch/Fun/funniest.html

Re:"Essentially" the same data?

[Anonymous Coward]

"A lot of office is loaded on startup."

I hear this argument a *lot* around here -- and I'm sorry to say, but it's complete Linux-apologist bullshit. No one has (as of yet) actually cited any evidence of this statement. All that they do is perpetuate it.

In an effort to discredit said Linux apologists, I actually performed a simple test:

1. Installed Windows XP SP2 in Virtual PC
2. After a clean reboot, I checked memory usage and all DLLs loaded in memory
3. Installed Office 2003
4. Re-did the checks

Conclusion? Nothing was being loaded. Office was not being "cached" for faster loading.

Re:"Essentially" the same data?

[Diego and Aline]

The professional statistics community uses Stata. Stop being so pushy about open source.

Re:"Essentially" the same data?

[joeljkp]

How's SPSS doing in popularity?

Re:"Essentially" the same data?

[Diego and Aline]

SPSS is more used in the social sciences -- sociology, psychology, etc. Another piece of statistics software I see a lot in the academia -- sp. in econometrics -- is RATS.

Re:"Essentially" the same data?

[Xabraxas]

I'm not sure about actual usage statistics but R is better than Stata.

Re:"Essentially" the same data?

[Javaman59]

Thanks! Well spotted!

Re:"Essentially" the same data?

[shintaro]

>they've just gotten better at hiding the fact that you need to load the entire office suite to load one application.

And I thought MS has the monopoly on doing things the dumb way. Why don't we call it as we see it? Loading the entire suite just cause I need to use Calc is pretty darn stupid.

On another note, OO 2.0 is bug-ridden. Had to return to 1.1.x after 10 minutes of using Calc because of multiple bugs (user problem). The main gripe(amongst other bugs) is this - Try cutting a small block of text in Calc 2.0 and see how long it takes, on my rig it takes 3-5 secs. Surely its not ready for prime time?

Re:"Essentially" the same data?

[letxa2000]

My wife took a college class on Office 2003, and it was obvious that the authors not only wrote the book in Word so they get points for eating their own dog food, but had split each chapter into a doc; so it appears that even the experts have trouble getting around this.

Heheh, I don't doubt it.

Actually, I've been meaning to investigate if there's some way in OO to put each chapter in its own document and then "compile" them into a single PDF at the end; of course, the trick is that the index, table of contents, and page numbering throughout the resulting PDF has to be seemless.

If there's some way to do this easily, I'd really like to do it. It's not because Writer has any trouble handling my 348-page book, it's just that it'd be a lot easier for me to have them separated into smaller logical units. When I'm working on chapter 14, I'm working on only chapter 14 99% of the time and all that other bulk just makes the scrollbar annoyingly small (i.e., dragging the scrollbar a little moves you 10 or 20 pages in your document).

And if the above is possible, it'd sure be great if there was a way to tell Writer that you want each new file (i.e. each chapter) to begin on a right-facing page.

If I could get these two things working, I'd be a very happy camper.

Re:"Essentially" the same data?

[Anonymous Coward]

So, then it is poorly designed. What is your point.

Re:"Essentially" the same data?

[Martin Blank]

I simply replicated the requested benchmarks. If you'd like to assist in coming up with what you believe to be a fair comparison, then by all means, let me know and I'll assist with it.

As for Windows pre-loading DLLs, I saw no immediate evidence of this, but when I get to work tomorrow, I'll have a look using ListDLLs and PsList after booting and before loading anything from Microsoft. I'll then publish results here.

Re:"Essentially" the same data?

[Hanzie]

Quoth the poster:

"most apps that originated within Microsoft tend to use purely descriptive names... I never did figure out Bob, though.)"

Perhaps if you visualize a corpse floating in the water...

Re:"Essentially" the same data?

[Anonymous Coward]

What those office slaves don't understand is Relational Database is built on RELATIONSHIP. That is, you don't repeat fields in every single table if it is not for eliminating excessive joins.
I inherited some Access (I hate Access in my guts!) codes and databases from previous coworkers. It is ugly, excessive large and slow design because they put the same freaking long text field into every table that needs it. Surely those people must know Excel better than Access (or Foxpro)

Re:"Essentially" the same data?

[fingusernames]

Heh. Just a few hours ago I had Excel 2003 SP2 crash two times (and nicely offer to restart) while diagnosing what turns out to be a known Excel bug. It was popping up a "the file may have changed, save a copy or overwrite?" dialog, which is apparently due to some lame time-stamp comparison method used to resolve contention, and which is apparently fixed in SP3. If any of you Samba users have experienced it, it appears that the underlying cause is that *nix has more resolution in timestamps than Windows expects. The non-SP3 fix is to mess with some registry key called, and I kid you not, QFE_Saskatchewan. See MS KB article 324491.

Larry

Re:"Essentially" the same data?

[jambarama]

I'm not saying that parts of office are preloaded. I'm saying office uses the same libraries that windows does. Installing office doesn't make more libraries preload as you demonstrated. Office uses a lot of the windows OS in it's own programs. This isn't bad; it is a great way to do things. In the same way koffice uses a lot of the kde libraries that are already loaded when you are in KDE.

Now if you get office to run in wine as fast as in Windows then maybe you'd have a case where office is just faster. But running it on a clean windows install doesn't disprove what I am saying.

Re:"Essentially" the same data?

[OhioJoe]

"OO.org took half the time to load that Excel did and took up just over half the space for the files."

I wonder what Vince Calloway is doing to get that kind of file opening rate.

(5 points if you know what I am talking about)
(10 points if you don't care)

OJ

Re:"Essentially" the same data?

[Anonymous Coward]

OO is probably slow because:
it's cross platform and has many transitional libraries (which take time to load, consume RAM). Excel, well, uses windows components some of which are already loaded in RAM.
And there are other windows specific optimisations which can help...

Gnumeric is best of open source world, anyway.

Re:"Essentially" the same data?

[Mikeytsi]

I thought RATS were more heavily used in laboratory science,......

Re:"Essentially" the same data?

[Anonymous Coward]

I think that they would've gone for Microsoft Sheet myself.

Re:"Essentially" the same data?

[Decker-Mage]

Among the statistical computing community it is still taught and rather popular or at least that is my experience. I started with SPSS back in the '70's (and used it for my first contract job) and was rather surprised to find it still taught, recommended, and in use by business and government when I went back to the university for another degree in the '90's. SAS is another popular package especially among the life sciences set that I sometimes work with. Still, given the fact that many of the high-end spreadsheet tools (Excel, etc.) and their capabilities, I've seen a reduction in demand for my services and utilization of these packages except for research, high-end enterprise, and government needs. That should be no surprise to anyone given the migration of capabilities down to the user level.

Re:"Essentially" the same data?

[rosciol]

I agree with you for the most part: people have a tendency to overuse Excel and they sometimes created bloated spreadsheets that would have been better utilized as databases. But I think we should be careful to criticize spreadsheets for the right reasons. Making spreadsheets that are painless to update is entirely possible, and doesn't require someone to be a wizard. In addition, using spreadsheets for financial forecasting and planning is intelligent.

Complicated reporting scenarios is one of the things that Excel handles best, because structuring database reports that involve any form of calculation or subtotals can be a real nightmare, especially for someone who isn't a SQL hacker. Subtotals, lists, names, formulas; all of these things can be set up to update automatically if you spend enough time with the spreadsheet.

What spreadsheets should not be used for is to update, store, or track, relational data. If the data in question is not relational, then moving to a database is the completely wrong idea. Then you have opposite problem, where you're using a database like a spreadsheet (which I've also seen far to often). In that case you're trading up in complexity and down in functionality since you're not using the new functionality that databases are providing you to any real extent.

In the case of reporting, however, Excel walks circles around any database application out there. There's a reason that a lot of the top packages provide ways to export data to a spreadsheet. What people do is use the database to structure the right data in the right way (the right use for a database) and then use Excel to manipulate that data into a fancy report that does more than just simply list things (the right way to use Excel).

Re:"Essentially" the same data?

[taylormc]

On Windows NT/2000/XP, try resetting the I/O memory buffer size to a sensible value (considering that when loading, OOo is rather I/O-bound), using the method given below (lifted from www.windowsnetworking.com). I find that a value of 8MB for my XP system (256MB) results in a significant improvement in startup and document load times.

------------
IoPageLockLimit controls the size of memory buffers for I/O devices. The default minimizes RAM usage. An I/O intensive system could benefit from larger buffer sizes. Caution: setting this parm too high can result in slower performance. Set it in increments and see how it effects your system. Windows NT / W2K / XP.

Hive: HKEY_LOCAL_MACHINE\System\CurrentControlSet\Contro l\Session Manager
Key: Memory Management
Name: IoPageLockLimit
Type: REG_DWORD
Value: 0x00000000 512KB
Value: 0x00100000 1MB
Value: 0x00200000 2MB
Value: 0x00400000 4MB
Value: 0x00800000 8MB

Re:"Essentially" the same data?

[Anonymous Coward]

I don't really care if OOO is slower than its MS equivelant. The real point is that it is an open standard and 'free'. I've been converting all my documents from MS and Corel to OOO, faster or not.

Gnumeric is a good program, but no version of it for Windows, that I know of. The great thing about OOO is that there's a version for pretty much every OS out there. It's good to know that I can type out a document in Writer while using my Linux box then send the document to my buddy who has OOO instaled on his Windows box.

Re:"Essentially" the same data?

[kimvette]

I submitted a defect report to the OOo team nearly a year ago about an Excel spreadsheet which OOo took 43 minutes to open but Microsoft Excel could open in under ten seconds - under wine! The spreadsheet came from The Commonwealth of Massachusetts. In Windows, it still opened in under ten seconds even with Norton AntiVirus scans.

The next month I received a spreadsheet which took well over two HOURS for OOo to open, but that same spreadsheet opened in Excel (via wine) in under 30 seconds, and under Windows even with Norton AntiVirus it took well under one minute. These were by no means large; the first was about 1200 rows in each of three worksheets, and the second was about 2500 rows in each of three worksheets.

When the feedback came back, they marked the priority down and indicated that they are more focused on adding new functionality since it is more interesting than fixing minor bugs. WTF? IMHO this is a fatal defect for all intents and purposes. Were I the release engineer or QA director for OOo/Star Office, I would not let the product go out as a final release in that condition since it is a serious hindrance to corporate adoptation.

With that said, I use OOo far more than I use Microsoft Office because I hate Microsoft's current Anti-Customer "every customer is a crook" stance, however I cannot fault Microsoft where they get things right. Microsoft Office rocks! Microsoft Exchange rocks! OOo is functionally a good replacement for M$ Office, but performance sucks when it comes to I/O. Likewise, Postfix is great and all, but when it comes to groupware, there really are no (open source) groupware packages that come even close to Microsoft Exchange's and Outlook's combined feature sets; this is coming from both a user feature/experience perspective and a maintenance perspective. Sure, Exchange backups may suck, but its maintenance tools are excellent in terms of maintaining back end integrity (providing you keep up with it), plus Exchange handles attachments extremely intelligently on the back end. Why do I mention this? Because I am biased against Microsoft and I want to make that clear so my post is taken in the correct context.

If you check the OOo bug tracker, you will see many, many reports of the very same defect I reported (the one cited above) and in each and every case this fatal defect has either been closed or the priority has been bumped down to "we'll fix bugs when Hell freezes over because fixing a broken architecture is boring, I'd rather add more bells and whistles and maybe an Easter egg or three"

I hope the OOo team gets their act together, because both OOo and StarOffice have incredible potential to truly compete with and possibly even overtake Microsoft Office in the future, but there are some critical issues (broken I/O) that they really need to address. I'd rather they address the issues now than wait for Massachusetts to deploy the suite only to find that with real-world Commonwealth data, the suite costs more in terms of wasted time than the Microsoft licenses would have cost. This is where Microsoft FUD would actually be factual, and then Microsoft can say "See, we told you so!" and come out smelling like roses.

Re:"Essentially" the same data?

[nlvp]

If Microsoft would quit loading all the query and PivotTable type features into Excel and concentrate on being a better spreadsheet than it is already, I wouldn't have any complaints. Even so, I suppose I grudgingly have to say that Excel is probably the least-offensive MS program out there.

I really disagree with this - Don't get me wrong, I love databases, have been a MySQL programmer for years and have built entire payroll analysis and calculation systems in Access, so I've no problem with using them, but they're not as rapidly manipulatable as an excel pivottable.

I work in finance for a very large company, and most of the data I obtain comes from SAP downloads in list form. I don't know if you use SAP, but the general gist of the thing is that while it's amazing at handling financial data because it was built for that purpose, it's really dreadful when it comes to rapid querying and manipulating data at speeds that allow for proper analysis.

Downloading all of the records in a cost centre for an entire year into Excel, pivoting it, and dragging the pivot-table components around allows for an extremely intuitive and almost instantaneous analysis of data. The alternatives are to use the SAP interface - it's slow and lacks functionality, or to use a second database to manipulate the data, which will take longer than the 60 seconds it takes me to pivot 20000 lines of ledger entries to pick apart a certain group of expenditures.

What's more, financial analysts tend to walk through the door with a knowledge of Pivottables from their previous job - we have to train them on SAP, and even those that have used Access in the past don't know how to do the same thing in Access as they can with pivottables, I've used access and MySQL for 10 years, give or take, and I can't manipulate a list of records as fast in those as I can in Excel.

Excel is an amazing tool, and the pivottable functionality is one of the most powerful features of it. I would find it very very hard to do my job as fast as I do it were this functionality to be removed.

Other features of Excel are also extremely useful - macros and protection allow us to consolidate data obtained through custom-built templates and collected from various parts of the business (which are then uploaded into a database once we get it), reporting spreadsheets have database queries built into them and standard report formats that allow users to interface with databases without needing days of training, because they never have to deal with the DB front end.

Excel has years of development under its belt, and apart from the occasional crash that leaves me cursing Microsoft and walking around the car park in an attempt to calm myself down, I can't think of a single other spreadsheet or data management tool that works so well, fast or intuitively with mid-sized data sets (i.e. less than 50000 records).

Databases are important when you're going to need the relational functionality you get form linking tables - which is why the raw finance data has to be held in SAP/Oracle.

I hate to sing Microsoft's praises, but I see no viable alternative to Excel that wouldn't result in a significant loss of productivity in the finance function, and I think this is easily extended to any business function that manipulates changing data sets intensively under time pressure. It's just built too much of a head start (in terms of development and iterations) on every other available program out there.

Re:"Essentially" the same data?

[Martin Blank]

Nothing from Office loads prior to launching any Office application, and that's done manually. The only references to "Office" that I found were from the OpenOffice quickloader files that were autostarted. There was no reference to anything in the Office path, no reference to mso*.dll, nothing that looked like it came from Microsoft Office in any way. If you want, I'll e-mail the result file.

Re:"Essentially" the same data?

[Kelson]

Now if they were willing to unbundle things more, you could choose to install only the parts of the application that you actually needed to use, which would mean you'd install a Microsoft Piece of Sheet.

Re:"Essentially" the same data?

[MrResistor]

While I think your post is very funny, and I love a MP reference as much as anyone, you seem to have missed something very important: the problem has nothing to do with Open Source! Let's take another look at that snippet you quoted, and see if we can make things a bit more clear, shall we?

"It is a known problem that OOo takes a while to start. Staroffice (at the point when Sun bought it) was made by a German company. Most of the internal functions are named in german, and use abbreviations that are not obvious."

So, hopefully now it is plain that this code was not made by some German guy working for free in his spare time, but was in fact written by a german guy getting paid by a company to write code, all of which was then bought by Sun, and then at some later point opened up to the world. The problem would still exist, and be at least as bad, if Sun had never opened up the code (probably worse, as it being open source at least allows for more eyes to look at the problem).

In fact, had it been a FOSS project from the get-go, this problem wouldn't exist. There's strong incentive when creating open source code to make it readable to others. Developers really don't like to work with code they can't make sense of, and in the FOSS world that means either your submissions get rejected, or no developers are attracted to your project. Thus, had it started out as FOSS, it either wouldn't have this problem, or the project would have died out years ago.

Clearly, no such incentive existed within the proprietary software company where Staroffice was originally produced. So, while you apparently were hoping to poke holes in the open source methodology, you have in fact pointed out one of the reasonss proprietary software development is seriously flawed: it often encourages bad code to be released to the public (in order to meet shipment, slopiness because "no one will ever see it", or even as intentional obfuscation).

Re:"Essentially" the same data?

[budgenator]

if there is a way I don't know it, and the reason that it was appearent that the book was done in Word was because the author's didn't either, 4 sections, word, excel, powerpoint and access, each individual page numbered, and I suspect that the chapters were numbered begining at 99 type of a thing. When they edited a page I'm sure that it made a mess of the numbering; Word and OO are word-processors, not document processors. When I'm doing something serious i.e. more than a page or two it's LaTeX for me.

I found it interest that the book was a Shelly and Cashman series book, those two taught me Fortran, Cobol and RPG II many moons ago!

Re:"Essentially" the same data?

[budgenator]

Even when I use windows, and start to get a mozilla, a couple instances of Wordpad, a command prompt for LaTeX and the DVI viewer all going at the same time, the window-space feels cramped and uncomfortable and it's hard to switch between apps without cycling through ones I don't want.

Re:"Essentially" the same data?

[Anonymous Coward]

Yup, the propensity to page out both program code and data equally without discriminating between the two often makes interactive responsiveness on my fast Linux systems seem less than on my ancient Solaris box.

Re:"Essentially" the same data?

[mikefe]

Try setting /proc/sys/vm/swapiness between 0 and 25.

Re:"Essentially" the same data?

[morgan_greywolf]

Actually, cutting and pasting a block of text within Calc 2.0 happens instantaneously for me. Also cutting and pasting a block of text from OOo Writer to OOo Calc 2.0 happens instantaneously for me. Cutting and pasting your post from Firefox to OOo Writer 2.0 happens instantaneously. What took time for me was cutting and pasting your post from Firefox to OOo Calc 2.0. That took closer to 3-5 seconds. Not sure if that's a 'bug' or not, though, because Calc has features such that if tabular data is sitting in the clipboard from other applications, it will format it into multiple cells automatically. Copying a smaller chunk of text (like a single number) , was near-instantaneous. I think probably that code is a bit slow. But the amount of times I copy and paste from a web browser to a Calc is quite negilibigle, so I can live with it.

MS Office set the benchmark for bloat

[Anonymous Coward]

So if you can do worse, your software is considered mega-bloated.

I know the problem!

[Anonymous Coward]

Java.

Re:I know the problem!

[tweedledopey]

I thought Java was the answer.

Re:I know the problem!

[beheaderaswp]

Could be...

I think it was Corel Draw that was released as a Java port a bunch of years ago. That was a bloat and a half from the standards at that time (I think I had 64 MB of RAM on a Mac 7100.

Java... is bloat. But somehow... I still love it.

Re:I know the problem!

[(A)*(B)!0_-]

You love Java yet obviously know nothing about it.

OpenOffice isn't even written in Java!

Re:I know the problem!

[Anonymous Coward]

only if the question was "What is the most hyped, under performing programming language?:"

Re:I know the problem!

[beheaderaswp]

Noo.... It hooks into it for a GREAT many things.

Re:I know the problem!

[eelke_klein]

I think it was Corel Draw that was released as a Java port a bunch of years ago.

You've got it partly right, it was Corel but the application was WordPerfect.

Re:I know the problem!

[Anonymous Coward]

.NET!

Re:I know the problem!

[(A)*(B)!0_-]

: Written in C++ and with documented APIs licensed under the LGPL open-source license, OpenOffice.org allows any knowledgeable developer to benefit from the source. And, because the native file format for OpenOffice.org is the vendor-independent OpenDocument open standard, interoperability is easy, making future development and adoption more certain.

Any other thoughts you have?

Re:I know the problem!

[Anonymous Coward]

Yeah.

Close your tags.

Testament to Open Source Software Developers

[Thanatopsis]

When attempting to replicate one of the biggest bloatware software packages out there, that they make a version even bigger and bloatier!

Re:Testament to Open Source Software Developers

[Anonymous Coward]

> When attempting to replicate one of the biggest bloatware software packages out there, that they make a version even bigger and bloatier!

OSS developers are like starving artists. It's not enough to recreate a painting of Elvis for free, you splash on a little black velvet baby!

Re:Testament to Open Source Software Developers

[Megane]

I would like to take this opportunity to coin the term "bloat-compatible".

A quick check with Google shows only one hit, on a page full of Baynesian Babble.

Re:Testament to Open Source Software Developers

[thesnarky1]

One might say they beat Microsoft at their own game!

Re:Testament to Open Source Software Developers

[mforbes]

I like it! Incidentally, until Google indexes this thread, the term "bloat compatible" is and will be a google whack blatt. I've tried to find one of those several times and never succeeded, so I want to offer you my congratulations for finding one without even trying!

Re:Testament to Open Source Software Developers

[xs650]

It look like the student has surpassed the teacher.

Re:Testament to Open Source Software Developers

[Anonymous Coward]

it's not a googlewhack. it contains punctuation (the -).

Re:Testament to Open Source Software Developers

[susano_otter]

Sadly, when the Teacher is a Sith Lord, the success of the student is to be dreaded, not anticipated.

How many slashdotters today are feeling betrayed, upon discovering that their Open Source Anakin turns out to be just another Bloatware Apprentice of Darth Sidious?

Re:Testament to Open Source Software Developers

[Megane]

it's not a googlewhack. it contains punctuation (the -).

But Google ignores punctuation!

And I'd like to point out that even though I misspelled it, "Bayesian Babble" also only has exactly one hit, also in a page of, what else, Bayesian Babble.

Re:Testament to Open Source Software Developers

[Megane]

Er. I just checked the definition of googlewhack, and it's not a googlewhack because I used quotes. Both times. (Hey, don't look at me, I never claimed it was, it was that other guy over there!)

Re:Testament to Open Source Software Developers

[demachina]

It would be interesting to know if someone has investigated using the symbol hiding capabilities in the newer versions of gcc to eliminate some of the shared object related bloat that most probably afflicts OpenOffice. When you use shared objects for everything every function name gets put in the dynamic symbol table by default. The only ones that actually need to be there are the ones called from the main program and other shared objects. All of the functions and global data that are only referenced by other code in the same shared object don't need to be in the dynamic symbol table or linked at run-time. Windows has used explicit exporting of symbols from the dawn of time, you can explicitly hide or export symbols in newer version of gcc, 3.4 in particular. I think KDE takes advantage of it on gcc 3.4 compiles.

You can look at the dynamic symbols that ARE loaded when the shared object loads with something like:

objdump -T /opt/OpenOffice.org/program/*.so

The bloat is especially accute in C++ code because the mangled function names can be quite long.

All those symbol names are loaded and scanned to do run-time link the shared objects, it causes slowness at startup which OpenOffice certainly has and you take a big memory hit for stuff that is not useful code.

Manually keeping track of which symbols need to be exported and which are not is a pain, and is a pain in Windows DLL's. You would almost be better off on something as big as OpenOffice to write scripts to process objdump output and figure out which symbols are actually be called outside the shared object and need to be in the dynamic symbol table.

On the other hand its kind of good discipline to create an a clean and disciplined API for each shared object which defines the public interface to the shared object. It helps improve modularity, reusability, testability and discipline in general and eliminate bloat when you realize that in fact nothing is actually calling dead code.

Re:Testament to Open Source Software Developers

[nietsch]

is it a true whack though? you use quotes to delimit the results. If you omit the results you get a lot more.
(BTW the bayesian babble wordlist page has already been slashdotted. Nice)

Re:Testament to Open Source Software Developers

[JumperCables233]

You know what fascinates me? The fact that people complain to no end about how bloated Office is, but if a single feature doesn't seem to be there they will scream to high heaven that the software is incomplete and inadequate. The bloat comes from the amount of stuff that Office can do. Granted, most of us only use about 10% of it, but if we wanted to use the other 90%, at least it's there.

Re:Testament to Open Source Software Developers

[GloomE]

True, by why do we have to pay for the resources used by this other 90% when we're not actually using them?

Re:Testament to Open Source Software Developers

[rtb61]

Don't be fooled, the only true comparisson is OpenOffice.org on Linux versus the non-standards compliant microsoft versions (talk about giving the big bucks, advertising spending monopoly the home ground advantage). Just how dumb would you have to be to not realise that there are always going to be performance issues in ensuring cross platform capability (Yeah I know journalism as marketing. I still laugh at the supposed tech journalists who are mentally incapable of installing Linux and complain about their own ignorance as if it is fault of the operating system). Open office on windows is just a transitional product to help people in the move from software exploitation to software freedom ;-).

Re:Testament to Open Source Software Developers

[demachina]

Uhhh, I'm not comparing the two, I seldom ever run Office. I just notice it take FOREVER to launch OpenOffice on a beefy Athlon with a gig of RAM. I pity anyone using it on a slow machine without a lot of RAM. It doesn't help I'm loading it on KDE so there is massive GUI toolkit duplication.

Sounds like you are just being defensive and an apologist. Fact is OpenOffice COULD use a lot of bloat reduction and startup speed improvement. I was thinking about taking a turn at it but it is a complete pain to build especially native under x86_64 on AMD. Someone needs to standardize that bloody build on configure and make, so more developers will pick up working on it.

Re:Testament to Open Source Software Developers

[Zixia]

Neither 'Bayesian Babble' nor 'bloat-compatible' are Googlewhacks. First, the two words should not be in quotes; they should be separate words that are included anywhere on the page idependently, not together. Second, a one-hit result doesn't count if the page in question is a dictionary or random list of words.

Consider the Source

[mfh]

Consider that Intel owns a big chunk of CNET and then you see a possible conflict of interest brewing over an article possibly designed to sink Open Office. Now consider the author, George Ou, who has also posted such titles as, Is the Honeymoon with Firefox Over?

Seeing a bit of a pattern forming.

Re:Consider the Source

[RPoet]

This is no bad thing, however. An article like this, pointing out that feature X of OpenOffice.org is n times slower than on Microsoft Office, will only trigger the OOo hackers to optimize and improve. So, in a sense, Ou's effort against OOo will "backlash".

Re:Consider the Source

[slavemowgli]

I'm not sure I can see the conflict of interest here. Does Intel have an office suite of their own they're trying to sell? Or did they merge with Microsoft recently? :)

Re:Consider the Source

[just_another_sean]

I always thought the industry abbreviation "Wintel" says it all. But then again the tinfoil on my hat just gets thicker every year.

Re:Consider the Source

[jsebrech]

What about the source? Ad hominem attacks are a logical fallacy. Who wrote the article should not have bearing on judging the validity of the article.

Now, what about the actual statements made in the article. Does anyone have anything to say about those?

Re:Consider the Source

[shbazjinkens]

I always thought the industry abbreviation "Wintel" says it all. But then again the tinfoil on my hat just gets thicker every year.

Well, saying MacOSXtel would just be too cumbersome and forward looking.

Re:Consider the Source

[gowen]

Indeed. Intel should be pushing OpenOffice, because nothing makes me aware of how much I need to upgrade my processor like starting OpenOffice.

Re:Consider the Source

[GileadGreene]

Linux runs on Intel too. And OOo runs on Windows or Linux. I doubt Intel cares what sofwtare is running on what OS, so long as everything is "Intel Inside®".

Re:Consider the Source

[jdclucidly]

I can't believe this got modded +5; moderators, shame on you. This post is the oldest logical fallacy in the book: Argumentum ad Hominem . Rather than attacking the source, you're supposed to attack the argument! And quite frankly, the tests look rock solid and statistically accurate. If you can't raise the level of the argument, just don't say anything at all. You make the rest of us look bad.

Your two digit ID be damned...

Re:Consider the Source

[Anonymous Coward]

What about the source? Ad hominem attacks are a logical fallacy. Who wrote the article should not have bearing on judging the validity of the article.

You clearly don't know what an ad hominem attack is. The ad hominem fallacy is where you say "Ou is an idiot, and I have reason to believe he's also homosexual. Therefore his article is rubbish". That is indeed a logical fallacy and an invalid argument.

On the other hand, to say "Ou has a well-documented history of writing negative articles on the subject of open-source software" is to state a fact, not to make an attack; and to continue, "therefore it is likely that his approach to the subject will be biased, his evidence selective, and his conclusions unreliable", is perfectly reasonable.

To be perfectly blunt, the provenance of an article is significant. If Linus Torvalds says "Linux is better than Windows", that means very little: of course he thinks that, and nobody really thinks twice when he says so. But if Bill Gates were to say the same thing, then it would be an incredibly significant statement, and people hearing it would immediately put great trust in those words: if Bill Gates says the competition is better, it must be really good!

Similarly, if an OpenOffice.org developer were to announce that their software was, in fact, not as good as MS Office, then that would be a significant announcement that should be given much credence. But when Ou, who has a long and easily verifiable history of writing articles that disparage open-source software, says the same thing, his words should be taken with a generous pinch of salt.

That's not an ad-hominem fallacy. It's called "critical thinking".

Re:Consider the Source

[nine-times]

Except, um.. you have Linux, BSD, and soon enough, you'll also have OSX. Intel actually supports Linux quite a bit, soooo...

Re:Consider the Source

[da_matta]

And even if it would be the case that George Ou is "anti-OpenOffice/Firefox or pro-MS", it doesn't make his arguments or point any less valid. He might be subjective, but he has argumented his case relatively well and backed it with data.

PS. Anyone here who thinks is not guilty of this, start throwing rocks.

Re:Consider the Source

[kayak334]

Couple those facts with the additional fact that pretty much everything he says about OO.org is true, and what do you get?

A worthless post.

Re:Consider the Source

[joebp]

The pattern you see is "open-source software is not yet as good as its closed-source equivalent."

Re:Consider the Source

[ArsonSmith]

Why resort to facts and logic when we can just degrade the authors character, motives and possible agenda. When A 300lbs person tells me I'm fat it automaticly makes me skinners because he is obviously much fatter than I am.

That "Pot and Kettle" story is BS. Just because they are both black doesn't make the information any less. Just because a 300lbs person tells me I'm fat doesn't mean I can't stand to lose 30lbs.

Re:Consider the Source

[Delphiki]

If only someone would bite the bullet and port OS X or Linux to an Intel platform maybe Intel wouldn't be spreading all this propaganda.

Re:Consider the Source

[Anonymous Coward]

All that "Wintel" refers to is the fact that Windows is the operatins system, and Intel is the microprocessor. It was never ment to imply that they are the same corporation or even share the same interests.

I've even once heard an Intel speaker lament that they are the "second part" of Wintel, instead of the first.

Re:Consider the Source

[Stupendoussteve]

Last time I visited Intel (their Oregon campus closest to Nike - don't remember the actual name), they had a whole bunch of computers in the cages that were in various states of Red Hat, Fedora, SuSE and Debian installs. In the cages, I didn't see anything running Windows except for the company PC's that the engineers did their actual work on (which there weren't very many compared to the others). The Linux stuff was for testing.

They were installing it on workstations and some of the other stuff, like blade servers.

Re:Consider the Source

[renderhead]

If Linus Torvalds says "Linux is better than Windows", that means very little: of course he thinks that, and nobody really thinks twice when he says so. But if Bill Gates were to say the same thing, then it would be an incredibly significant statement...

This only holds true for vague, relative and subjective statements like the one in your example.

A better example would be if Linus Torvalds said "I ran a test that demonstrates Linux booting 25% faster than Windows on the same hardware," and Bill Gates responded with "My tests show that Windows boots 15% faster than Linux on the same hardware."

Yes, both speakers have motives that are worth questioning But the proper response is not to dismiss both claims because of the speakers' biases but rather to take a closer look at their methods and their results. If you find problems with their tests, then you can dismiss their results. If not, then you must accept the results and attempt to reconcile them with any conflicting data you have encountered.

It doesn't matter how reputable the source is. You should always check their research before you accept their claims as conclusive.

Re:Consider the Source

[fimbulvetr]

You must be comparing this to how people complained about kde being slow and bloated and the kde people devels turned it around so it ran faster and took less memory. Oh wait...that hasn't happened yet.

Re:Consider the Source

[diegocgteleline.es]

Indeed - every article pointing out that firefox is not having lots of vulnerabilities lately (it's false! false! Secunia is lying dammit!) and that openoffice is bloated (how they dare! Everybody I know says openoffice is lean and fast!) is manipulated.

Re:Consider the Source

[Nijika]

Ou is more a naysayer than a crypto-shill for MS. I watched him beat XML with a crowbar the other day.

Re:Consider the Source

[Total_Wimp]

Then there's the old standby, "there's no such thing as bad publicity."

I know a lot of people who'd prefer bloated, slow and free(as in beer) to less bloated, faster, but $379.95 at newegg.com. Many of those people don't even know there's an alternative to MS Office. I hope they read C/Net.

TW

Re:Consider the Source

[Mr. Shiny And New]

Actually, each version of KDE has been faster than previous versions on the same hardware. I can attest to that, I noticed major speedups when I upgraded from version 2 to version 3.

Re:Consider the Source

[xappax]

It's true that the source of information doesn't affect whether or not the information is true, however it does affect how trustworthy the information is.

In this case, since the author obviously has a history of spreading an anti-open source message, the trustworthiness of his information is pretty low. Now, if you're presented with a factual argument by someone you distrust, the appropriate thing to do is to test their claims yourself. In many cases, further testing or investigation reveals the claims of untrustworthy people to be inaccurate, however this is not a given.

In this case, although I haven't repeated his exact experiment, I strongly suspect that his claims are accurate despite his obviously slanted reason for presenting them. I also suspect that the reason many people are so eager to assume that his untrustworthiness makes him wrong is the realization that performing an objective inquiry will probably confirm his argument.

Re:Consider the Source

[protoshoggoth]

Yes, yes. This would all be much more interesting were it not for the fact that however biased he is and whatever he has said in the past, his statements on this matter appear (to me, who hates office as much as any right-thinking person would) to be pretty much spot-on.

Re:Consider the Source

[V_Pundit]

mod parent up.

Re:Consider the Source

[fonetik]

Wouldn't it be better to read the article, no matter what the viewpoint is, and apply a bit of scientific method to it? Rather than dismiss it out of hand because of who owns what and what he wrote previously? How would you like it if after making an effort to be fair, and did the experiments to see where performance or user experieince differs in the two packages, only to have it posted on slashdot and dismissed on the spot because 'Slashdot is biased towards the OSS community, so they probably lied.'?

It's the same thing as if you are on the left and you see something on FOX News, you can dismiss it with no further thought and you might be correct to do so. A lot of things have been shown to be wildly inaccurate at best or frequent lies by omission on that channel. But I would think that true "critical thinking" would be figuring that out for yourself, or at least watching it and confirming or denying the group opinion for yourself.

It would be a shame to overlook something that may need to be addressed just because it came from a source that you don't like. If Open Office really is that slow, it's a valid mitigating factor in not using it. If it's misleading, find out why. But I think it's just as foolish to ignore something because you don't like who told you it, much less, to expect everyone else to ignore it too.

Re:Consider the Source

[sidb]

This post is the oldest logical fallacy in the book: Argumentum ad Hominem

Ad hominem (latin: argument to the man) is perhaps the second oldest logical fallacy. The oldest, and still the best, will always be argumentam ad baculum (latin: argument with a stick). Regrettably, the advent of the internet as a medium for argument has rendered this time-honored technique somewhat less practical than it used to be, but I'm sure a creative debator could still find a way to employ it.

Re:Consider the Source

[rpdillon]

Actually, yes, it has. One of the biggest areas of development in KDE has been (and continues to be) cleaning up code and streamlining the system. I run KDE on 5 of my personal computers, and I can tell you: later version systematically run cleaner and faster than earlier versions, expecially 2.x.

Re:Consider the Source

[jacekm]

Well, if you own a stopwatch, you can quickly prove or disprove the author claim. All it takes is a computer that has Windows XP and Linux installed with dual boot and both office applications. If you don't like message don't shoot the meesenger unless you can proove that the the messenger is actually lying. Since when the test financed by some corporation must be always biased ? I know, that some liberals and greenes always suspect "big corporations" of doing nothing else, but this is in most cases totally baseless. And finally why is it in the Intel interest to cheat in order to protect Microsoft ? Intel will probably do as good as today even if everybody would switch to Linux. JAM

Re:Consider the Source

[Anonymous Coward]

It's not called "critical thinking." It called "appeal to authority" which is a also a logical fallacy. Every example you list is a logical fallacy and the only part that contains any facts is your definition of ad hominem. This is not "critical thinking."

Re:Consider the Source

[SoccerManUNLV]

Well, he does post his finding, and everyone i've seen who tried to replicate the same tests have failed to even come close to his posted results. You can try and see what you come up with as well if you go here: http://blogs.zdnet.com/Ou/?p=119 and download what he used. I tried these tests, and came close to his findings once, and i'm running a much slower machine. I'd have to say he's "fixing" his results in favor of office, but thats only my opinion, you can do the test yourself and see how they vary.

Re:Consider the Source

[jdclucidly]

+1 funny

Re:Consider the Source

[Anonymous Coward]

At one time I would spell check my comments, but grammer/spelling nazis are just too funny. I missed them.

It's spelled grammar not grammer!

Re:Consider the Source

[Anonymous Coward]

No. No, no no no no.

The whole reason behind ad hominem attacks being wrong is that they don't point to any actual proof of anything regarding the argument. Just because someone is an idiot doesn't mean what they are saying is incorrect. Just because someone has shown a bias toward a viewpoint, doesn't mean that they're wrong, either.

In that sense, saying that "Ou has a history of writing negative articles on the subject of open-source software" IS, in fact, attacking the messenger, not the argument presented. It is absolutely an ad hom attack. Sure, such knowledge might lead you to carefully examine Ou's claims (which might, in turn, lead you to discover real evidence against Ou's assertions), but it is *NOT* evidence about the claims themselves.

I'm critical of your thinking.

Parent post is completely wrong and makes the "Insightful" tag cry. Mod it down!

Re:Consider the Source

[cagle_.25]

On the other hand, to say "Ou has a well-documented history of writing negative articles on the subject of open-source software" is to state a fact, not to make an attack; and to continue, "therefore it is likely that his approach to the subject will be biased, his evidence selective, and his conclusions unreliable", is perfectly reasonable.

No, it's not perfectly reasonable. Your argument confuses point-of-view with improper methodology.

Only if his articles showed previous evidence of bias, selective evidence, and unreliable conclusions, could one conlude that his articles will likely do so in the future. His ability to use facts responsibly has nothing to do with which side he's on.

That said, it would be reasonable to say "Ou is making an argument about the bloat of Open Office. Therefore, I should find out how the OO programmers justify the size of OO, since presumably they are not idiots." That is, Ou has showed that OO is large, but not unjustifiably so.

Oh, and we should check his facts ... but we should do that regardless of which side he's on.That is "critical thinking."

Re:Consider the Source

[Mysticalfruit]

Moore's Law: The speed of processors will double every 18 months.
Gate's Law: The speed of software will halve every 18 months.

Re:Consider the Source

[Arandir]

When you make an argument based on WHO THE PROPONENT IS, then you are making an ad hominem argument.

Re:Consider the Source

[Anonymous Coward]

But the proper response is not to dismiss both claims because of the speakers' biases but rather to take a closer look at their methods and their results.

The issue is not only if the claims are true. That you can solve by testing or reasoning.

At issue is also if the claims are even worth considering; or testing. Ad Hominem attacks are often not making the point that the specific claims are wrong, but that the specific claims are not worth considering.

(If that seems unreasonable, consider this claim from an AC: Java is 44% slower then Python in making SSL decriptions on alpha hardware. Do you want to investigate that? Java vs. Python speeds may be important, but that claim is'nt, and the source matters.)

Re:Consider the Source

[SeattleGameboy]

Gosh, this is one of the stupidest thread I have ever seen.

I can understand yours and the parent poster's reservation about the article's authors bias IF the article was just stating his opinions, which would be impossible to prove false, but...

BUT HE ACTUALLY POSTED NUMBERS!!!

You don't like the results? Just do the tests yourself. We are not talking about global warming here. Just load both apps, creates some files and see how long it takes to open it.

If you think he is biased and is wrong, just go ahead and PROVE IT!!!

Gosh, don't we have enough people right now (re: media) who confuse facts with opinions? I think we can do better than that hear at Slashdot.

Re:Consider the Source

[civilizedINTENSITY]

Wintel vs. Intdows? Sounds like "Ain't Those". I've *always* heard it suggested that that "wintel" refers to a collusion between Microsoft and Intel. Googling for "wintel collusion" finds many who define "wintel" in terms of collusion, suggesting that MS Bloat sells Intel's newer, faster chips. I've heard this, like, almost forever...

Re:Consider the Source

[doublestar]

Paul Allen ( Microsoft Co OFunder ) has a large investment in CNET Works .. he invested $5 million US in 1994 .. this is the parent of ZDNET

Re:Consider the Source

[jmorris42]

> It doesn't matter how reputable the source is. You should always check their research
> before you accept their claims as conclusive.

Not in the real world. In the real world there are various constraints, mainly time. There simply isn't enough time to track down and verify every statement everyone makes. Therefore it is very helpful to be able to cull out the known cranks, crackpots and axe grinders.

That said, Captain Obvious over at ZDNet isn't exactly leaking classified information when he says OO.o is a bloated C++ (and since C++ isn't bad enough, Sun is adding extra Java suckiness!) horror and slow as molasses in January. Being pro Open Source doesn't mean we have to pretend OO.o doesn't need some serious performance tuning. The miracle is they shipped the feature set in 2.0 at all, now they need to stop adding features for awhile and make them work right.

Re:Consider the Source

[Anonymous Coward]

Before you talk about a conflict of interest, you have to evaluate the truthfulness of the article. I think it's at least semi-truthful.

I have to say that I have experienced speed and performance (read resource) issues with Open Office on the aging Windows 2K machine that I use in my daily work. This machine has a 1300 MHz cpu and 785 MB of RAM and was recently rebuilt and redeployed especially for my use by an apparently competent engineer. Migration to Linux in this environment is not a practical alternative at present, so I can't address that.

"My" machine, to be sure, has a number of performance issues, some related to the network (NetWare in use; many files on networked drives) and some to the software that the local shop uses for desktop management and security monitoring. Other problems are probably related to drivers, motherboard, graphics card, and other issues whereof I wot not--there are some fundamental problems with MS. apps hanging as well, particularly Visio. Tinkering with all of that is off limits to me.

In general, however, while the functionality of O-O is superb, its performance in this environment is less satisfactory than that of the Bill Gates stuff, while the file-level compatibility, though better than that of Corel Office, is only so-so. How much of this is due to restrictions imposed by the O/S I can't say. The net effect, however, is to make it impossible for me to recommend O-O to my internal customers. They will have no sympathy with my desire to run screaming from Word/Excel/Access etc. They have a lot of work to do and want to take their Office tools for granted.

I had hoped to recommend the O-O DocBook features as a way to begin migrating our documentation efforts from MS Word templates and doc files on a shared drive to some form of single source publication with the primary venue being a content-managed Web. While initial results with the eDE command-line tools were promising, in the end I was unable to devise a process smooth enough to be used by anyone but me.

Thus, in this environment, O-O offers only an alternative but not-noticeably-better Office setup that works for many people less well than Microsoft AND that offers no piece of a practical alternative to the endless pileup of document files on hard drives that has replaced the paper mountains of yesteryear. It can be used, but will not be perceived as competitive except in cases where the lack of a purchase price is decisive; e.g., when you are starting a company, department, or agency from scratch with all-new computers, perhaps using Linux.

The bottom line is this: Lots of people who deal with lots of documents every day in business, government, research institutions, etc. are stuck with file-based, MS Office-based, proprietary solutions to the problem. The high-end "Enterprise" doc. origination and management tools fail to meet this need for many, if not most, organizations. Manually attaching PDFs to Web pages and publishing really technical docs.via frame-based HTML image-map drilldowns are still bleeding-edge in this marketplace. To win its niche decisively, Open Office has to both perform better than Word & Co. as a standalone word processor, etc., and provide at least the biggest user-end piece of a superior technological solution for single-source document origin.

O-O is of incredible value to people who need a high-quality Office tool for no money, especially when they are starting from scratch with new computers. But to conquer the world, I think O-O will have to offer both superior end-user performance (which is the main reason for insisting on superior resource utilization) AND the user-end key piece of next-generation text-document publication (and pub. for presentations and spreadsheets, etc.). Otherwise, potential users may just buy SharePoint and wait for standalone Office tools to become obsolete.

Re:Consider the Source

[khallow]

Wouldn't it be better to read the article, no matter what the viewpoint is, and apply a bit of scientific method to it? Rather than dismiss it out of hand because of who owns what and what he wrote previously? How would you like it if after making an effort to be fair, and did the experiments to see where performance or user experieince differs in the two packages, only to have it posted on slashdot and dismissed on the spot because 'Slashdot is biased towards the OSS community, so they probably lied.'?

I don't know about this situation, but I'd say "no" in general. The problem is that you have a finite amount of time. Considering the provenance of a bit of information is useful in determining whether you should spend more time in examining the information.

Re:Consider the Source

[Knetzar]

These days it seems that X.0 adds features, while X.1 adds performance and stability.
So I'll wait until OO.o 2.1.

Re:Consider the Source

[generic-man]

Consider that OpenOffice.org is descended from StarOffice, a commercial application that literally tried to mimic every feature in Microsoft Office and did so quite poorly. Until the open source community, bound by no "roadmap" or "release schedule," can clean up OpenOffice.org's code, OpenOffice.org will be amazingly slow.

On Linux, it's the most feature-complete office suite around. On Windows or Mac, it's still too early to consider OpenOffice.org superior to the competition.

Re:Consider the Source

[StopSayingYouSir]

On the other hand, to say "Ou has a well-documented history of writing negative articles on the subject of open-source software" is to state a fact, not to make an attack; and to continue, "therefore it is likely that his approach to the subject will be biased, his evidence selective, and his conclusions unreliable", is perfectly reasonable.

No. You've simply retreated into fuzzy inductive territory. This does not make it any less of a fallacy; it simply shows that there is sometimes a difference between formal logic and what we might call common sense.

Re:Consider the Source

[LiquidCoooled]

Its not just gates.
Its practically every single programmer out there.

The majority of packages nowadays run awfully on former top of the range processors.

Its the nature of developers in general, I certainly don't code as close to the metal as I used to - it just takes far too long for deminishing returns.

Re:Consider the Source

[dtfinch]

I think Gate's law is every 12 months.

Re:Consider the Source

[InvalidError]

Moore's Law: transistor count doubles every ~18 months.

Up to about three years ago, this translated in a comparable increase in system performance. If you look at system performance now compared to back when Northwood/HT was introduced, the only thing that sort-of-doubled performance since then is dual-core. Add marginal clock increases and infrastructure upgrades and you get a ~3X boost over the last three years rather than 4X. They can keep doubling transistor count by doubling caches and core counts but beyond 2MB caches and quad-core, we're deep in the land of diminishing returns.

Now, it takes nearly triple the transistors to double performance, bringing the performance-doubling cycle closer to 24 months... and even longer in the not-so-distant future.

Re:Consider the Source

[Burz]

That's what I've noticed as well. KDE has been improving its performance.

However OOo doesn't really use KDE, or QT or GTK. It has its own widget set and abstraction layer, which means its an OS within an office suite. Oddly enough, it still cannot be easily ported to a platform like Mac OS X without first having to install X11.

Re:Consider the Source

[penguinoid]

I can't believe this got modded +5; moderators, shame on you. This post is the oldest logical fallacy in the book: Argumentum ad Hominem . Rather than attacking the source, you're supposed to attack the argument! And quite frankly, the tests look rock solid and statistically accurate.

From Wikipedia:
Ad hominem is fallacious when applied to deduction, and not the evidence (or premise) of an argument. Evidence may be doubted or rejected based on the source for reasons of credibility, but to doubt or reject a deduction based on the source is the ad hominem fallacy.

Premises discrediting the person can exist in valid arguments, when the person being criticized is the sole source for a piece of evidence used in one of his arguments (emphasis mine)

So if this is the guy who did the tests, an "ad hominem" argument can be to call him a rotten liar and wonder whether his data is accurate. Any argument based on such data relies on doubtful premises and is therefore invalid.

Re:Consider the Source

[jalefkowit]

The oldest, and still the best, will always be argumentam ad baculum...

Not to be confused with Argumentum ad Baculum: the technique of answering any challenge with the assertion that "Enterprise" wasn't a crappy show.

Re:Consider the Source

[MBGMorden]

The problem is most "people" either get Office with their computer or install using a buddy's CD (don't come up posting "Well I don't!", because I'm not accusing any particular person, just saying that the average home user does). The people who pay for Office are large commercial entities, and to them, the cost of Office isn't all that high, especially when you can always purchase discounted upgrade licenses.

I can honestly say that if I want to get my employer to switch to OpenOffice, or any other software program, it's going to have to be BETTER, not cheaper. I do what I can to promote using open source software within our organization, but it's rare that cost is even brought up when trying to convince the powers that be.

Re:Consider the Source

[dave420]

Great! That solved the problem. :) OpenOffice is streamlined again! yay! :)

Re:Consider the Source

[Billly Gates]

Actually if you google to sysinternals (I think thats the site... dont have time to look it up) and profile openoffice you will see how it handles threading and processing.

I read somewhere that openoffice performance can be improved by 20x alone by proper threading and object utilization in the win32 version. For example office only loads dlls it needs and unloads them when done to save ram.

It has nothing to do with language and more about design. Staroffice which openoffice came from was a unix program and many unix programs do not use so's like windows ones use dlls. Its a mess.

Re:Consider the Source

[nietsch]

Yes, he posted statistics.
Are you aware what can be done with statistics? Are you aware that the document he tested with might not relate to reality? Could it be that he picked that form just to prove his bias?

Re:Consider the Source

[Eccles]

I've been using OO Impress for the past few days now, and I'm perfectly satisfied with its performance. C++ may be bloated as a language, but C++ compilers produce mighty fast code. (Isn't most of Microsoft's stuff written in that same C++?) In my experience, OO may need some tuning, but maybe not to handle the relatively casual user. (On the other hand, Word is often painfully slow at loading images within DOC files, and could definitely use some performance tweaking.)

Re:Consider the Source

[SeattleGameboy]

WTF??? Did you even read the article???

The CPU comparison was done with NO DATA. With just apps launching and doing nothing else. If you still think he is doing something with statistics, try it yourself. Just open the apps and check the available resources.

The comparison between opening documents is as trasparent as they get. He even POSTED THE SAMPLE FILES HE TESTED WITH!

SXC - http://www.lanarchitect.net/Examples/200264-l.sxc

XML - http://www.lanarchitect.net/Examples/200264-l.zip

Read the files. Find out if the documents are fixed. If you think he did, provide documents that OO will open faster than Office.

C'mon. Conspiracy charges may work with assasinations, but this so easy to test yourself, you have no excuse to charge bias without providing some facts.

Are you REALLY that lazy? OR whatever you don't agree with, you just label "biased"? May be you should go work for FOXNews...

Re:Consider the Source

[Maxo-Texas]

What he is saying is...
Not
A better example would be if Linus Torvalds said "I ran a test that demonstrates Linux booting 25% faster than Windows on the same hardware," and Bill Gates responded with "My tests show that Windows boots 15% faster than Linux on the same hardware."

But
A better example would be if Linus Torvalds said "I ran a test and it demonstrates Windows booting 25% faster than Linux on the same hardware," and Bill Gates responded with "I hate to say it but my tests show that Linux boots 15% faster than Windows on the same hardware."

Similarly
If George Ou said, "I tested Openoffice and it blew Word away." Folks would believe him because he supports windows and disses opensource and such a statement is probably honest and after he checked everything. For George to say Windows stuff is good is almost a non-statement because he has no credibility any more. He spent it all by being relentlessly pro-windows and anti-linux / open source over the course of several years.

On the general subject, as someone else said:
OO really is slower and uses more memory because it does things in software that Word and Excel can ask the operating system for. That's a minus for OO.
Word and Excel really are proprietary and require you to pay more money every year or two for an upgrade. That's a minus for them.
Word and Excel really do lock you into windows while OO runs on most OS. Windows intends to go to a subscription model where you have to pay money every month to get current patches. Another minus for them.

Re:Consider the Source

[Total_Wimp]

I can honestly say that if I want to get my employer to switch to OpenOffice, or any other software program, it's going to have to be BETTER, not cheaper. I do what I can to promote using open source software within our organization, but it's rare that cost is even brought up when trying to convince the powers that be.

"Better" is a great selling point, but don't underestimate cost. My company went with JawsPDF because we just couldn't swallow the cost of Acrobat on 700 desks.

We're not going to switch to OO.o any time soon because we've already paid for MS Office and there's no end-user demand for OO.o at this point. But we strongly evaluate the cost of every project and we do sometimes pick "good enough" products based solely on a favorable price. I doubt we're the only company that makes their decisions this way either.

TW

Re:Consider the Source

[nietsch]

sure it was with 'no data'. That 3.6 Mb is just white noise then, I suppose? FYI: it is 16 sheets of cut & paste 'data' with 16282 rows per sheet. You still want to argue that this is not a doctered set?
And no, I cannot compare the two, the MS product does not run on linux. (the file took ~8 min to load on my system)
My argument is stil valid: he made it look like it was an objective comparison. But only one data point (one file to load) and no explanation why or how this file was chosen is not carefull research. This is FUD/mudslinging.

Re:Consider the Source

[Anonymous Coward]

No, it is when you attack the person as a way to distroy his arguments. It is perfectly reasonable to look at a series of articles that a person has written and point out bias in those articles to prove that more than likely any article this person writes in the future can be discarded as more of the same. This is not an "attack on the man."

--

From the web site http://www.infidels.org/news/atheism/logic.html#ho minem:

Argumentum ad hominem

Argumentum ad hominem literally means "argument directed at the man"; there are two varieties.

The first is the abusive form. If you refuse to accept a statement, and justify your refusal by criticizing the person who made the statement, then you are guilty of abusive argumentum ad hominem. For example:

        "You claim that atheists can be moral -- yet I happen to know that you abandoned your wife and children."

This is a fallacy because the truth of an assertion doesn't depend on the virtues of the person asserting it. A less blatant argumentum ad hominem is to reject a proposition based on the fact that it was also asserted by some other easily criticized person. For example:

        "Therefore we should close down the church? Hitler and Stalin would have agreed with you."

A second form of argumentum ad hominem is to try and persuade someone to accept a statement you make, by referring to that person's particular circumstances. For example:

        "Therefore it is perfectly acceptable to kill animals for food. I hope you won't argue otherwise, given that you're quite happy to wear leather shoes."

This is known as circumstantial argumentum ad hominem. The fallacy can also be used as an excuse to reject a particular conclusion. For example:

        "Of course you'd argue that positive discrimination is a bad thing. You're white."

This particular form of Argumentum ad Hominem, when you allege that someone is rationalizing a conclusion for selfish reasons, is also known as "poisoning the well."

It's not always invalid to refer to the circumstances of an individual who is making a claim. If someone is a known perjurer or liar, that fact will reduce their credibility as a witness. It won't, however, prove that their testimony is false in this case. It also won't alter the soundness of any logical arguments they may make.

--

So therefore, it would be wrong to say that he is gay, therefore ignore his argument, or that he is a windows user, so therefore ignore his argument. But saying that he has a long history of bias against anything open source, presenting his series of articles that are all biased, and he works long hours to find bizzare test cases that show excel as being better than OO, therefore what he says in the future will more than likely also be biased is a perfectly reasonable argument and is NOT an attack against the man. Of course he may be right in this one case, but who wants to waste the time to even check, given his track record. Certainly I have better things to do.

Re:Consider the Source

[Arandir]

Argumentum ad hominem literally means "argument directed at the man"

But isn't that exactly what is happening here? The rest of your post is interesting, but just because a particular argument doesn't fit its examples precisely and exactly doesn't mean that it is not an ad hominem argument.

When a person's ideas are rejected specifically because of who the person is, then an ad hominem argument is being made. An untrustworthy/biased/politically-incorrect man's should always be double checked, but to simply reject an *idea* because of who its proponent is, is wrong.

The OP post was correct, in that we SHOULD consider the author. But not as an argument to reject the premise (OO.org is bloated), but as a warning to get a second opinion or to investigate for yourself.

Re:Consider the Source

[Dispenser of Mayhem]

This has been announced by Apple recently: "Apple to Use Intel Microprocessors Beginning in 2006" http://www.apple.com/pr/library/2005/jun/06intel.h tml

wow

[krelyk]

Talk about stating the obvious... have you never tried to compile an app written by Sun before? The bloat just slaps you in the face repeatedly until you submit

I've used OpenOffice

[Deathly809]

And from what I have seen it does run a bit slow. If you try to open a file from the web might as well go get a drink. Maybe its just my computer?

Re:I've used OpenOffice

[Anonymous Coward]

What tit moderated that flamebait. This is a guys opinion, disagreeing does not make it flamebait.

Re:I've used OpenOffice

[Deathly809]

Im flamebait? Better than jailbait atleast...

How much difference between Java and C++?

[plover]

How much of this slowness is the application's fault vs. this being a giant Java app running in a JRE? AFAIK, MS Office 2003 is still a suite of (mostly) C++ applications, and isn't running in .NET yet.

Re:How much difference between Java and C++?

[BenjyD]

Openoffice.org is a C++ app. It uses java for some scripting, but everything else is C++.

Re:How much difference between Java and C++?

[InThane]

AFAIK, most of OO is NOT using Java - the only part of OO in Java is the database manager, and that's only for the JDBC connectivity.

Re:How much difference between Java and C++?

[Glock27]

How much of this slowness is the application's fault vs. this being a giant Java app running in a JRE?

OpenOffice is *not* written in Java. It'd most likely work better if it were. ;-) It is written in C++. I wonder if there'd be much of a speedup compiling it with the Intel compiler....

It does have some Java functionality, which is why a JRE is required. IIRC, gcj is the JRE most often used, which might impact interpreted Java performance. Gcj has a slow interpreter, though I think the most recent version has an optional JITC.

Re:How much difference between Java and C++?

[jdeisenberg]

OpenOffice.org is not a "giant Java app." Some of the features require Java, but the main app is written in C++. See the build instructions.

Re:How much difference between Java and C++?

[plover]

Thanks for the clarification. I had always believed it was primarily a Java app (especially considering the strong backing by Sun.)

I wish I could retract my previous comment.

Re:How much difference between Java and C++?

[MrNemesis]

I'd chance my arm and say a fair bit.

I made the mistake of opting for x86-64 Gentoo for one of my desktop boxes ("upgrading" it to 32bit this weekend), meaning I have to use the 32bit precompiled OpenOffice binaries. But these need hooking into a 32bit JRE which x86-64 Gentoo doesn't have, since making 32bit apps available through Portage is seemingly something that Gentoo Won't Do Because You Should Be Happy With 64bit. So whenever you start OOo it spends about a minute looking for a JVM (and failing) before you can do anything. I could have manually installed Sun's 32bit JRE, but I can't be bothered.

Disable Java in the options and it starts in 1-2 seconds on the same machine.

By way of comparison, I tried the same trick on my 32bit box (similar spec but with slower HDD's) and OOo was as snappy as hell and opened like the proverbial soil off a shovel.

If there's any functionality I miss through disabling Java, I haven't encountered any yet. And please note I'm not saying that Java is slow to execute (it isn't), it's just appallingly slow to load.

Re:How much difference between Java and C++?

[Anonymous Coward]

I decided to conduct my own brain surgery, and encountered problems, I can't imagen why...

Re:How much difference between Java and C++?

[confusion here]

Strange, I'm running 64-bit Gentoo and I have not experienced any issues with oo.org. When I first installed I did `emerge openoffice-bin` and it has been working perfectly ever since. Yes, it's 32-bit, but the integration is seamless.

Re:How much difference between Java and C++?

[Anonymous Coward]

Gotta love the "+5 Insightful" on the OP. ROFL!

Re:How much difference between Java and C++?

[Bradee-oh!]

Disable Java in the options and it starts in 1-2 seconds on the same machine.

Somewhat off topic but pertinent ENOUGH... Good God man! Thank you! The Java tab in the options dialog was incredibly easy to find but for some reason I just breezed right over it. Unclicking that little devil's box just dropped my start time from 15-20 seconds to 1. I know it likely has nothing to do with the working data that this "benchmark" tested, but it sure shows how good an idea it would be to transition the Java dependency on over to native code.

Re:How much difference between Java and C++?

[oyenstikker]

but I can't be bothered.

But you can be bothered to post your complaints on slashdot? It doesn't take that long to install another jre.

Re:How much difference between Java and C++?

[tedgyz]

How much of this slowness is the application's fault vs. this being a giant Java app running in a JRE? AFAIK, MS Office 2003 is still a suite of (mostly) C++ applications, and isn't running in .NET yet.

C'mon! Do we have to blame Java every time there is a performance issue? As someone pointed out, the core pieces of Open Office are C++. When will we stop fanning the flames of Java performance myths? Prolly when we stop bitching about Microsoft. :-)

Re:How much difference between Java and C++?

[Haeleth]

IIRC, gcj is the JRE most often used, which might impact interpreted Java performance. Gcj has a slow interpreter, though I think the most recent version has an optional JITC.

Um, I believe the JRE most often used is Sun's. In the open-source world, Kaffe is quite popular. GCJ is a GCC front-end that compiles Java to native code; a slow interpreter is optionally included in libgcj, but it's not built by default and it isn't the primary way of running GCJ code.

Re:How much difference between Java and C++?

[johansalk]

Hi. You're absolutely right! Mod parent up, I just tried this.

Re:How much difference between Java and C++?

[Anonymous Coward]

Debian's OpenOffice version is pretty quick, relatively. It's certainly much quicker then the old version. It takes 5-7 seconds to load the first time. I close it and open it again and it loads in a second or less. Entirely unscientific, of course.

2.0ghz Althon XP, 1gig ram, 80gig WD 8meg 'caviar' disk. Debian Sid.

I think it's one of those cases were apps works much better in Linux then Windows. In windows your comparing a highly optimized and integrated C++ application that uses as much as the system libraries (pre-loaded with windows explorer and everything) versus a very third party application that rely's on it's own toolkit and java stuff.

Similar things that I've found are nicer in Linux then in Windows is things like Gimp, Blender, Firefox, and a handfull of others that I can't think of.

Re:How much difference between Java and C++?

[Anonymous Coward]

I donno why they're considered myths. You can show me as many laboratory tests as you'd like, but until they match up with real world experiences . . .

Re:How much difference between Java and C++?

[ArgieNomad]

Don't retract it, you're getting some modlove just because your question "looks like a good one".

Re:How much difference between Java and C++?

[Anonymous Coward]

Awesome! I can't find this pref in NeoOffice/J on the Mac though :(

Re:How much difference between Java and C++?

[Glock27]

Um, I believe the JRE most often used is Sun's.

Actually, you're right, given that OO is mostly used on Windows. I'm sure the original article was testing under Windows as well.

I was actually referring to the Linux version, which on RH and I assume other distributions, uses gcj so as to have a "free as in speech" JRE.

Sorry for the mixup!

Re:How much difference between Java and C++?

[CyricZ]

Often times Java is responsible for slow GUI applications. That's not a myth, as much as Sun would like you to think it is. Try even a mid-sized application like Columba. You'll notice that it feels far slower than Thunderbird, or Seamonkey, or Outlook Express, even when using a modern JVM.

In this case, as previous posts here have suggested, the use of a JRE for scripting capabilities is resulting in long startup times, even if most of OO.o is written in C++. Disabling Java support resulted in vastly improved startup times. That does lead us to believe that Java is somewhat responsible for some of the problems that have been noted.

Re:How much difference between Java and C++?

[matth]

Sorry I'm not seeing this tab under options.. I just see a list of stuff on the left... where are you seeing it?

Re:How much difference between Java and C++?

[bruce_s_r]

Just be sure that the difference in startup speed is due to the java being disabled and not due to the fact that you just had OOo open a few seconds before. I just tried opening oowriter and it took about 30 seconds. Closing and opening again took about 1 second. It was so short a time I wouldn't have notice any difference if I then tried again with java disabled.

Re:How much difference between Java and C++?

[piquadratCH]

Unclicking that little devil's box just dropped my start time from 15-20 seconds to 1.

Make sure that this speedup of the second startup(without Java) isn't the result of all the caching after the first start (with java).
I tried it myself on my P4 1.8GHz: the first start (with java) took about 15s, subsequent starts (with java) took 7s. After disabling java, the startup time didn't improve. All times are handstoped with a wrist watch, so YMMV.

Re:How much difference between Java and C++?

[cnelzie]

If you close the "OpenOffice" quick starter application, after disabling Java (As I just did) you will indeed notice an increase in loadup speed from a "Cold start".

    It is impressive.

    Now, if for some reason, I need Java, I can always enable it later and then restart the application for the time that I will need Java.

Re:How much difference between Java and C++?

[Master of Transhuman]

Well, in this case, I don't think it's Linux vs Windows. I just did the same experiment on my Windows XP side using OpenOffice 2.0: first load using Quickstarter, 8 seconds. Second load, after turning off Java, and presumably relying on the Windows cache, one second. My CPU is the same as yours, 512MB RAM, 160GB WD HD.

I haven't tested any large spreadsheets for large memory usage, but I haven't noticed any particular slowness about OO. But then I don't use Excel or an Office stuff on a daily basis, so I can't really compare.

In response to the article, it should be noted that Microsoft's Vista is going to need 512MB of RAM just to run the OS, so I don't think Ou's in a position to complain about OSS "bloat."

Also, I'm not sure memory bloat is even relevant except to the degree that it perceptibly slows down the entire system. If you have enough memory to use the app effectively, then by definition it's irrelevant. Once you start running a ton of apps at once, both Linux and Windows will swap, and then the issue is how fast the user can multitask and how fast the hard drive is in swapping the apps back in.

In other words, less memory usage is a nice design goal, but how often is it a real issue? On servers running server apps it may be, but how many people really run OpenOffice from a server?

Considering that Windows XP has forty million lines of code, and Linux doesn't come anywhere near that, who is the purveyor of bloat?

Re:How much difference between Java and C++?

[level_headed_midwest]

I tried the disable Java checkbox and it seems to maybe take a second off of the load. Writer 2.0 loads on my 2.2 GHz P4-M machine in about 5 seconds now and so does the rest of the suite. It is much quicker than 1.1.4, which took about 15 seconds to load. MS Office under Windows takes about two to three seconds. A difference of a second or two is not all that big of a deal, at least to me. Sure, OOo might be a couple of seconds slower, but it's not that big of a deal. I save time using OOo on Linux because I don't have to restart Office or the whole computer when Office or Windows borks up something.

Re:How much difference between Java and C++?

[Glock27]

In this case, as previous posts here have suggested, the use of a JRE for scripting capabilities is resulting in long startup times, even if most of OO.o is written in C++.

No, the only post that seems to complain about long startup times is dealing with the case where there is NO JVM INSTALLED, but OO is looking for one.

Disabling Java support resulted in vastly improved startup times.

Since OO was no longer searching for a JVM that wasn't there in the first place...

That does lead us to believe that Java is somewhat responsible for some of the problems that have been noted.

Logic error, retrying....

Re:How much difference between Java and C++?

[CyricZ]

You admit that Java was causing the problems. Then you pretend you can't understand that fact. Like it or not, the problems were solved when Java was removed from the picture. Hence, Java is to blame for the problems.

Re:How much difference between Java and C++?

[MrNemesis]

The fact of the matter is, I think Gentoo is great. But I'm put off by their eternal reaction that everything should have a native 64bit version, and if it doesn't then stuff it. Getting 32bit apps to work nicely and seamlessly on Gentoo can be an exercise in pain - if you want to be able to play anything that relies on win32codecs (WMV, etc etc) you have to manually download and install an unsupported 3rd party ebuild of mPlayer. If you want something like Xine (like me), you have to build your own. Because of the way portage works, if you want a nice frontend instead of the appallingly awful GTK interface you have to compile your own from scratch.

Same thing for the JRE. I don't want to have to manually install anything when it's only the stubborness of the devs that keeps precompiled 32bit apps for AMD64 users out of portage. I don't want the hassle of updating it manually. As far as I can see, there's no technical reason why installing openoffice-bin with the java USE flag doesn't pull down the 32bit JRE as a dependency.

I'm complaining here because requests for features like this in the Gentoo forums seem to fall on deaf ears. It's the reason I'm switching back to a 32bit install - I like Gentoo, but I'm not prepared to put up with its incredible clunkiness WRT to mixing 32 and 64bit apps when distros like Fedora don't have the same problem.

Re:How much difference between Java and C++?

[MrNemesis]

I was wary of that too. The times I reported were done after just logging in after a reboot, so there shouldn't be anything cached. To make doubly sure, I turned off swap and still saw a colossal speed increase. This was on Linux and I don't have the quiskstarter installed.

It seems a significant portion of OOo's startup time was waiting for the JVM to load (since I don't have any other Java apps and rarely see/use it on web pages), and at the moment disabling it doesn't seem to give any less functionality. I imagine there may be problems with the new database stuff, but I use UNIX ODBC for my DB connection which bypasses Java.

Re:How much difference between Java and C++?

[Anonymous Coward]

Damn! Just disabled it myself. It boots in under a second now!

Re:How much difference between Java and C++?

[plover]

heh. modlove.

10% Troll
30% Overrated
20% Flamebait

Oh well, what's karma good for if not to get a bit of forgiveness for when you make a big, ugly mistake?

Re:How much difference between Java and C++?

[LnxAddct]

Fedora has kind of side stepped this problem with GCJ. All of OpenOffice's Java stuff runs natively, and despite the fact that GCJ doesn't optimize Java code as good as GCC optimizes C code, it still runs blazingly faster than on systems where a JVM is necessary for full functionality. Pretty nifty if you ask me.
Regards,
Steve

Re:How much difference between Java and C++?

[baadger]

It amuses me that on your machine has a OO startup time is "5-7 seconds". I just tried it on an old Windows 2000 machine with 128MB's of RAM with a PII 233 Mhz gut. It took 21 seconds from being clicked to being completely ready and disk silence.

These processes of course have a lot of influencing factors, but I'd say a 10 fold increase in the number of clock cycles (not forgetting new instruction sets and processing tricks) and 8x the (hopefully) available RAM, for only a 3 fold increase in application loading time, is a definate indication that the bottleneck lies in software.

I'm perfectly happy with office application performance on that box (yes of course I have another more duper machine), it's rather worrying how minimum specs have increased for a huge number of apps that IMHO, don't need the processing power. Perhaps it's just that programmers are becoming (by nurture or by nature?) sloppier or more reliant on huge (possibly unnecessary?) abstract layers of code.

P.S. For the record, MS Word 2003 starts in 7 secs. This means all the time you're using OO and i'm using Word, my P2 is just as satisfying as your much more beefy machine? Worrying indeed.

Re:How much difference between Java and C++?

[Anonymous Coward]

Hey, it works! OpenOffice becomes much faster!!!

THANKS!!! :-)

Re:How much difference between Java and C++?

[tedgyz]

I will not dispute what you are saying. I also can't quite understand why Java is dragging down OO.o. My general complaint and the use of the word myth is directed at the blanket statement that if a program uses Java, it is slow.

With that said, I believe a good example of a fast Java GUI application is Azureus. As bittorrent clients go, it is quite amazing.

Re:How much difference between Java and C++?

[Anonymous Coward]

Too bad.. bloat problem would be fixed if it was written entirely in Java. Garbage collecion is a wonderful thing.

Re:How much difference between Java and C++?

[Anonymous Coward]

You know, you could just grow a pair and install a 32bit JVM without the use of portage.

My main gripe with Gentoo is that people think its "hardcore" because everything is compiled from scratch, when in reality its one of the easiest distros to install. Everything is done for you.

I finally matured linuxly when I had to install radeon drivers from scratch (about 8 months ago when they were such a mess it almost wasnt worth it). Now just about everything seems easy by comparison.

Re:How much difference between Java and C++?

[aaronl]

Actually, just in casual use I have found two non-script places. The first is the well known database engine in Base. The second is the MS-XML filter.

I have the oddest suspicion that if they wrote the whole thing in the most appropriate single language, it would be faster. I *know* that MS-XML filter would be faster had they written it in C++ instead of Java.

People just need to get out of this habit of picking their favorite language or toolkit or Perl module instead of picking the most appropriate. Having to load a JRE for a few random things when you have a largely C++ project is a waste of resources. It was immature to do what they did, and the problem is too common. They didn't *finish* the product before they claimed it was release quality. They made it feature complete, they fixed many of the bugs, but they didn't keep going to the end. It's the same problem that Firefox and Thunderbird have, for example. You can't put together a product that does the same things as everyone else, but does them all slower, and expect people to not complain.

I still plan on replacing Office with OOo, but I would've liked for users to have seen an obvious improvment. Instead of noticing so many of the nicer features and methods, what they're going to see is that it takes a lot longer to start, longer to load files, longer to print, etc. The way that OOo has been handled bothers me quite a bit, as you can tell; enough that the two reasons left for moving to it are cost and the OpenDocument format.

Re:How much difference between Java and C++?

[Curl E]

Sloccount says:

Totals grouped by language (dominant language first): cpp: 4525856 89.71 java: 343291 6.80 ansic: 96307 1.91 perl: 48153 0.95 sh: 10484 0.21 yacc: 6785 0.13 cs: 5114 0.10 lex: 2458 0.05 python: 2354 0.05 asm: 2045 0.04 awk: 885 0.02 objc: 702 0.01 pascal: 397 0.01 csh: 272 0.01 php: 104 0.00 sed: 3 0.00

The first figure is the number of lines, the second is the percentage of the total lines. Sloccount also says Total Estimated Cost to Develop is $ 208,768,434. The output form the program is much more pretty than that above, but the lameness filter is preventing it's display. Is that filter preventing lameness of enforcing it?

Re:How much difference between Java and C++?

[CyricZ]

It does help that it uses SWT, rather than Swing. So you're getting the speed benefits of using native widgets, as well as the avoidance of the performance-awful Swing framework. Then again, Azureus isn't exactly a large application like OO.o.

Re:How much difference between Java and C++?

[nathanh]

Just be sure that the difference in startup speed is due to the java being disabled and not due to the fact that you just had OOo open a few seconds before. I just tried opening oowriter and it took about 30 seconds. Closing and opening again took about 1 second. It was so short a time I wouldn't have notice any difference if I then tried again with java disabled.

Seems easy enough to test. Load it once to cache then load it twice more. During each load I hit Ctrl-Q and measure the time from start to finish.

Cached w/ Java #1: real 0m14.333s
Cached w/ Java #2: real 0m14.607s

Same thing with Java disabled.

Cached w/out Java #1: real 0m9.002s
Cached w/out Java #2: real 0m8.982s

So there's a definite improvement. Not quite as good as 20 seconds -> 1 second though. I suspect the gp was seeing the benefit of the data being cached.

I'll be leaving it enabled on my machine. The startup time isn't really relevant as I tend to spend 5-6 hours inside Writer once it's started.

Re:How much difference between Java and C++?

[MrNemesis]

Are you using 1.x or 2.0? I don't recall having any problems with 1.x, but 2.0 is much more reliant on Java it seems, and leads to the atrociously slow startup times.

Re:How much difference between Java and C++?

[tedgyz]

Then again, Azureus isn't exactly a large application like OO.o.

Mayhap, that is the real issue. We all bitch about MS Office being bloated. OO.o is turning into a bloated alternative.

Re:How much difference between Java and C++?

[MrNemesis]

Sure, I could just install them manually - it's quite easy after all. Indeed, I've compiled and installed plenty of my own stuff already - lots of plugins for GIMP, Amarok and other misc utils for the most part.

The problem I have with going for a non-portage route for Java is that it's be a struggle to keep it updated. Since Java would be the only app on my system that could represent a mojor security problem if a bug was found, my days would be spent scanning Bugtraq to see when I should upgrade the JRE next. There's no technical reason portage can't do this for me, it's just the portage maintainers being stubborn.

In the meantime, I'm only gonna be using the AMD64 arch for servers where things like win32codecs and SSE/SSE2 optimisations don't make a whole lotta difference.

Re:How much difference between Java and C++?

[Glock27]

despite the fact that GCJ doesn't optimize Java code as good as GCC optimizes C code,

GCJ does just fine. On the one benchmark I really messed with, almabench, GCJ ended up within 5% of gcc with the right command line switches (-ffast-math, -funroll-loops, -fno-bounds-check IIRC). They both use the same backend optimizer, so that makes sense.

On the other hand, gcc isn't all that great at optimizations yet. Intel C was about twice as fast as gcc on that benchmark. Gcc 4.0 has potential, but isn't a big improvement - hopefully future versions will be.

(BTW, I don't think the gcj compiler is used by OO, just the interpreter gij...I'd like clarification on this.)

Re:How much difference between Java and C++?

[orasio]

I have the oddest suspicion that if they wrote the whole thing in the most appropriate single language, it would be faster. I *know* that MS-XML filter would be faster had they written it in C++ instead of Java.



Java is great for parsers. Debugging them is easier. When they fail, they throw an exception. When C++ programs fail, they can make vulnerabilities, sometimes.
It's much easier to have a robust parser in Java. A good parser, a fast parser, maybe it would take more work, but robustness is key.

Re:How much difference between Java and C++?

[aaronl]

It's an issue of the Java XML classes, not so much Java being slow in general. The need to do constant DOM traversal and such makes it sluggish.

While debugging and things may be easier in Java, they're certainly doable in C++. My point was simply that writing a project in a mixed pile of languages is not the best way. Java might be easier to debug, but 30 linked Java programs, a bunch of C++ programs, a few different APIs, and various differnt platforms to worry about is never easy to debug. Having the mix of languages just adds to the difficulty. Plus, you have to deal with even more IPC, and you have to deal with a software requirement that you can't distribute.

If it was all Java, then it wouldn't be as much a problem. You load the VM when you start it up, and everything is fine after that.

Re:How much difference between Java and C++?

[confusion here]

Ah, I'm using 1.x. I try to stick to amd64 rather than ~amd64 when I can, so that difference didn't even occur to me.

Re:How much difference between Java and C++?

[petermgreen]

i'm not sure on the details but i belive that gcj can be set up to act as a caching jit for gij then the cache can be preseeded so that known dependencies don't even have to be built on demand.

GUI

[kevin_conaway]

Could it be the GUI? Excel uses native widgets and I'm sure is heavily optimized towards MFC (after all, its their API!). I don't think OO has that luxury. I doubt thats the entire issue but it could partially explain it.

Re:GUI

[bheer]

Office uses zero MFC. Most of the older bits is Platform SDK C, and is a b*tch to maintain, and the newer parts are C++ *but not* MFC -- I understand the Office team has its own lightweight frameworks, similar to ATL.

Re:GUI

[Anonymous Coward]

If the analysis is accurate, does it really matter? The end user cares about performance - not the 'why' of performance.

Re:GUI

[mnmn]

Well they DO have other options. FLTK and QT both are close to native in speed of usage. (I havent checked what they use). I wonder if the GUI can be easily changed to some other API, so people could download the version thats fastest for them (OpenGL anyone?).

It doesnt seem to be the GUI thats slow. Once it starts, its fast enough. Its the loading part.

Re:GUI

[Anonymous Coward]

Excel doesn't use native widgets at all.

Re:GUI

[FidelCatsro]

I would be interested to see the test done on linux using Codeweavers crossover office .
(Well I don't know if the windows version of OpenOffice works on linux using Wine or such like ).
See if MS is up to some old tricks

Re:GUI

[alienw]

This would definitely be a factor if you were running it on, say, a 486. Try out the GIMP for Windows, there is no perceptible difference in GUI responsiveness, even though it uses GTK+ instead of the Windows API. I think the main problem with OpenOffice is that it's an ancient codebase and tries to do too much internally. Someone designing it today would probably use platform-specific features more actively instead of trying to make it look the same on every platform (which was the meaning of "portability" about 15 years ago). Not to mention, StarOffice was always a crappy, bloated product and OpenOffice isn't much better.

Re:GUI

[truthsearch]

They also have their own C++ compiler. Excel can't be compiled with Visual Studio.

Re:GUI

[electroniceric]

IIRC there was an article not too long ago that explained that the main difference between Office and OO is that Office makes extensive use of lazy loading, while OO essentially hammers through loading every library it may need, which not only thrashes the disk once on initial load, but again as you (likely) swap out memory pages. My recollection was that this lack of lazy loading had something to do with cross-platform compiling and linking issues, as well as MS having extensive resources to put into optimizing Office loading that OO did not. My understanding is that Sun hasn't exactly dumped developer time into OO, either, and I believe the focus of this release was compatibility and.

Typically people solve this problem by preloading a bunch of the relevant libraries at startup, a strategy both MS and OO attempt to employ (viz OfficeStartup and OO QuickStarter). I used to detest that, but if I had 1 or 2GB or RAM and wanted to rely on OO, I might not find it so bad. I think an interesting addition to this comparison would be to see how OO fared with QuickStarter enabled, and what drain that placed on the rest of the system. Likewise disabling the JVM loading.

Re:GUI

[bheer]

Old tricks indeed. The only 'trick' I see here incompetence -- on the part of whining developers who write cross-platform code and then complain that Microsoft's platform-optimized code runs faster than theirs.

Netscape complained for ages that Navigator couldn't be as fast as IE because of 'hidden DLLs'. Pooh. They (and the early Mozilla 5 team) never bothered to read MSDN -- DLL rebasing, page fault reduction: all of this is documented there, but you have to be willing to actually read about and use them.

Today, Firefox uses them and Firefox's load times are almost as good as IE's (if slightly less due to the XUL overhead). The Moox builds, which optimize more aggressively, start even better. Even the Seamonkey devs clued up, current Seamonkey builds start MUCH faster as a result.

Re:GUI

[silviuc]

I really can't say. Thing is, MS Office 2000 is snapier and loads faster (almost instantly) than OO 1.1.3 or OO 2.0 on my P II @ 333Mhz machine with 256 MB of RAM. Oh yeah, I run MS Office with Wine.

Re:GUI

[Scoth]

In my experience, Office seems to be one of the most common abusers of using non-native widgets. For example, even in Office 97 they reimplemented a lot of the window controls and menu handlers. If you ran it in NT3.51 you'd still get the Win95-style menus and widgets in the MDI interfaces. Office XP and 2k3 use the shiny shaded menus with pretty highlighting and such.

That said, it would be interesting to compare the speeds of the libraries used.

Re:GUI

[interiot]

Windows Vista also reserves a chunk of RAM and disk space so that Excel can have its own memory manager and swap file. It's thought this C++ (complexity++) will make it more difficult or impossible for MSOffice to run on emulators such as Wine.

Re:GUI

[GNUALMAFUERTE]

Off course, and not only the GUI, but many functionality built into m$ office is actually part of the OS, while OO is a complete suite, with all the code into the app itself. It's the same that happends with IE. IE loads faster than Firefox, because IE is allready loaded as part of the OS, with the actual explorer being just a frontend, if we preload Gecko, Mozilla would load faster (Konqueror is a good example of this).

Anyway, We use Free Software beacause it's Free (As in Freedom). If a certain app has technicall problems, we should be working to improve it, not write bad reviews about it.

Re:GUI

[CyricZ]

FLTK is hardly suitable for this scale of development. It's good for small apps, but it most likely does not scale to an application the size of OpenOffice.org. Last I knew, it still handled callbacks using function pointers. That's not a very clean design at all.

Qt is perhaps a better choice, from a technical standpoint. But then there are licensing issues that may need to be addressed, especially when it comes to Sun selling their modified version.

And no, OO.o uses their own GUI framework. Transitioning to a toolkit like Qt would most likely be quite difficult, time consuming, and perhaps without benefit in the end.

Re:GUI

[CyricZ]

Where do you get that Excel cannot be compiled with Visual Studio? Are/were you a Microsoft employee, who is/was on the Excel build team? If so, and your assertion is true, what compiler are they using instead?

Re:GUI

[shotfeel]

I'm not only wondering about the GUI, but the OS in general. How much of the code that Office apps use is preloaded as part of the OS? This not only effects how much RAM an application appears to use, it can also be a determining factor is speed benchmarks.

Re:GUI

[Anonymous Coward]

Wine
Is
Not an
Emulator.

Otherwise, and interesting post.

Re:GUI

[Tim Browse]

The only 'trick' I see here incompetence -- on the part of whining developers who write cross-platform code and then complain that Microsoft's platform-optimized code runs faster than theirs.

Even more so, when you consider this.

Re:GUI

[lseltzer]

I'm calling shenanigans. Where the hell did you hear this?

Re:GUI

[qbwiz]

Yes, because as we all know, Crossover Office is a magical fairy-tale that does not exist. Why would Vista reserve RAM and a swap file specifically for Excel? Can't Excel malloc and mmap on its own?

Re:GUI

[callipygian-showsyst]

There's NO MFC in Office. While I am a big fan of Microsoft, and make about half my income managing software projects that run on Microsoft OSs, I'd be the first to tell you that MFC was a failure. Very few people used it, inside of Microsoft or outside. Microsoft has deprecated it years ago.

Re:GUI

[DigitlDud]

Actually they DON'T use native widgets. Office has been getting into trouble over this for quite some time.

Re:GUI

[Anonymous Coward]

No. Office is always compiled with the latest released version of the VC++ compiler.

Re:GUI

[Anonymous Coward]

I heard gcc. No, wait that was just the voices in my head.

Re:GUI

[shmlco]

Having been in five different shops that used MFC to create Windows-based applications, I'd have to call that statement BS.

Re:GUI

[Anonymous Coward]

Then you must be using a completely different GTK from the rest of the world. Because GTK is significantly slower (we're talking second-long delays to pop up menus!) on Windows than, well, anything other than Java.

Re:GUI

[interiot]

LOL. Crap, I go to lunch, and come back to find out it's been modded "interesting". It was meant to be a joke. I don't know how much more I could have possibly exagerated it.

Vista will also require a special motherboard that allows Excel to run on a unique kind of CPU, separate from dual-core chips of today. New power supplies that output 17.23 volts will also be required. Buckle up, folks.

Re:GUI

[delphi329]

Office even has its own compiler and assembly code from 1988!

Re:GUI

[nasor]

My laptop has a 1.3GHz processor with 512 mb of ram (not top of the line any more, but still pretty damn far from a 486) and there is a very noticeable difference. I ended up uninstalling Open Office shortly after I got it and returning to MS Office. Files that would open virtually instantaneously with MS Office would commonly take 30+ seconds to open with Open Office.

Re:GUI

[Geoff NoNick]

Bingo - this is exactly the difference between OpenOffice and MS Office. But ultimately it's just a matter of exchanging the OS' virtual paging system for an internal system that loads and unloads as required. You'll still end up getting a fair amount of swapping in both cases, but the latter (which MS Office uses) makes it *appear* that there's less memory being used. In reality, most of OpenOffice's used memory is committed to the swap file anyway.

Re:GUI

[LordSah]

I'm a former Office Developer (now happily at a start-up). I can tell you with certainty that Office is compiled with the VC compiler. Not everyone in Office uses VS as their IDE, but it's compiled with VC.

Re:GUI

[truthsearch]

Well, all I can tell you is that I was told this by Office developers in the usenet newsgroups. Some of the changes they made to the compiler saw their way into the VC compiler.

Re:GUI

[LordSah]

I don't know the specifics of Excel, but Outlook does make use of MFC. It's a crazy mish-mash of MFC, with Win32 API windowing calls, and here and there is new UI technology. InfoPath and OneNote, both new products, are probably 100% new stuff. Word and Excel certainly have less MFC (and possibly none), because the codebase is so old. It's relatively rare that MS will re-write giant chunks of UI just to support a new technology. Why spend tons of dev time to make the app have the exact same set of features and appearance it had? They (usually) spend the dev dollars on improving features or writing new ones.

Re:GUI

[d_lesage]

Agreed. Four for four, in my experience.

Re:GUI

[Yaztromo]

My guess would be they might be using Intel's compiler (are they still calling it Proton?).

While I was working at IBM a couple of years back, that was the compiler of choice for the project I was working on (at least for when compiling for Windows), as it was found to have a significant performance advantage over any of Microsoft's compilers. It would make sense that Microsoft would see a similar performance advantage in their products.

Yaz.

Re:GUI

[LordSah]

How long ago was this? I worked there from 2001 to last year, and I think that may have been true a _long_ time ago (Word is over 20 years old). Office will dogfood unreleased VS compilers (and feedback does go into refinements for VS), but no one ever mentioned an Office-specific compiler.

Re:GUI

[interiot]

LOL, so now it gets modded all the way down because I didn't make it 100% clear it was a joke? In all other cases, at worst, people would have ignored it as a bad joke. But because it gathered far more attention than I intended, it gets dropped down to 0. How funny.

Re:GUI

[bani]

now happily at a start-up

working at microsoft sucked that badly?

Re:GUI

[ArmorFiend]

but if I had 1 or 2GB or RAM and wanted to rely on OO

We've got 3/8ths of a gig in the laptop upstairs. Though it is slower, we've switched it to OO over Abiword, abiword just being too wonky and cantankerous for general use anymore (recent problems include hanging when doing Print Preview and being out of Wifi range of the hub, and random rectangles inserted into PDF renderings of the document).

And on the downstairs workstation with a gig, OO is positively snappy! I recommend it to everyone. ^^

Re:GUI

[Anonymous Coward]

My machine at work has 1GB RAM and I do leave the quickstarter on all the time. I was stunned with the final version of OO 2 - the apps start almost instantly (1 sec) for a blank document, compared with about 10 seconds for 1.9.113 which I previously tried out.

Strangely the same thing on my home PC (also 1GB RAM and a faster CPU / hard disc / memory) takes about 3 seconds for Calc or Text.

Re:GUI

[bheer]

Joel Spolsky was, among other things, PM on Excel pre Office 95. http://www.joelonsoftware.com/articles/fog00000000 07.html and http://discuss.fogcreek.com/askjoel/default.asp?cm d=show&ixPost=4201 might be of interest.

Re:GUI

[LordSah]

Whether or not you're happy at Microsoft depends on a number of factors: your team, your boss, your product, etc. Working on a small team on a brand-new product is usually fun (you get a lot of impact on the product, lots of room for promotion). Working on an old product with a huge team with lots of bureaucracy is generally bad. The big benefits cuts and culture changes (no more foosball tables, way fewer morale evets, etc) in the last couple years have contributed to a general downtrend of morale at the company.

Is it (generally) a good place to work? Yeah, it is. Is it as good as Google, or MS of 6 years ago? Nope.

Re:GUI

[Jesus_666]

By the way, is there a way to somehow cut down on OOo RAM usage? I have an iBook with 256 megs of RAM (emergency purchase after emergency purchase leaves me perpetually short sixty bucks to put in a 512M module) and rely on NeoOffice/J to do office-y stuff. Needless to say, using the thing feels like, to paraphrase the game "Psychonauts", molasses going up a hill in january. With crutches.

Re:GUI

[dusanv]

FYI, completely different code bases I think.

Re:GUI

[Anonymous Coward]

It was before your time. Grep for 'compiler' on that page.

Re:GUI

[Tim Browse]

I very much doubt that. At one stage (Office v6), the Mac and Windows versions had basically exactly the same GUI code (as well as all the other code). That's changed (mainly due to the poor reception from Mac users of Office 6), but I find it hard to believe that MS completely re-implemented Word, Excel, etc, just for the Mac. That is a lot of code to write again (for no particularly good reason).

Re:GUI

[macshit]

Even more so, when you consider this.

My god... I thought that page was a joke until I looked at the URL!

Microsoft's entire web site has a sort of sickeningly sweet pandering-to-the-clueless-majority feel to it, but at least it's relatively low-key pandering. This subsite ("mactopia") on the other hand, is cringe-inducingly awful -- it's like the same dull web-designers that do MS's general site were trying to mimic Apple's style, but ended up producing a rather cruel mockery of it instead.

Re:GUI

[Breakfast Pants]

A long time ago. Someone already posted this link, but here you go: http://www.joelonsoftware.com/articles/fog00000000 07.html

Re:GUI

[BKX]

I think YOU must be using a different GTK than the rest of us. While several years that may have been true, today I use many, many GTK+ (2.6.9) apps on Windows and they all seem FASTER than their native Windows counterparts. OOo 2.0 loads faster than WordPerfect or MS Word ever did. Gaim does everything better. Gimp loads in only a second or two, whereas Photoshop tool much, much longer. Why do you try something less ancient before bitching about it.

Re:GUI

[lseltzer]

I apologize for contributing to the general environment of credulousness on /. If I hadn't already posted here I'd mod you "funny"

Re:GUI

[silverkniveshotmail.]

Didn't office start as mac-only though?

Re:GUI

[Tim Browse]

You're thinking of Excel, which launched on Mac OS. The first release of, e.g. Word was for DOS.

Debugging the Development Process by Steve Maguire has some interesting insight (mostly as asides or side-bars) into the cross-platform nature of Office, in particular Excel, because Maguire was one of the driving forces between getting the core functionality to be cross-platform. Before that there was a constant disparity in features between the Mac and Windows versions.

And here is another quite interesting history of Word, in case you think it's a latecomer to Windows.

Re:GUI

[silverkniveshotmail.]

from http://en.wikipedia.org/wiki/Microsoft_Office

Major Microsoft Windows versions

        * Office 3.0 (CD-ROM version: Word 2.0c, Excel 4.0a, PowerPoint 3.0, Mail) - released August 30, 1993
        * Office 4.0 (Word 6.0, Excel 4.0, PowerPoint 3.0) - released January 17, 1994
        * Office for NT 4.2 (Word 6.0 [32-bit, i386 and Alpha], Excel 5.0 [32-bit, i386 and Alpha], PowerPoint 4.0 [16-bit], "Microsoft Office Manager") - released July 3, 1994
        * Office 4.3 (The last 16-bit version; Word 6.0, Excel 5.0, PowerPoint 4.0 and in the pro version: Access 2.0) - released June 2, 1994
        * Office 7.0/'95 (Word '95, etc.) - released August 30, 1995
        * Office 8.0/'97 (Word '97, etc.) - released December 30, 1996 (was published on CD-ROM as well as on a set of 45 3½-inch floppy disks)
        * Office 9.0/2000 (Word 2000, etc.) - released January 27, 1999
        * Office 10.0/2002/XP (Word 2002, etc.) - released May 31, 2001
        * Office 11.0/2003 (Word 2003, etc.) - released November 17, 2003
        * Office 12.0/2006 - due to be released simultaneously, or near simultaneously with Windows Vista, Microsoft's next major consumer operating system.

There are variants of more recent versions such as Small Business Edition, Student and Teacher Edition, Professional Edition and Developer Edition with different collections of applications and pricing points.
[edit]

Apple Macintosh versions

        * Office 1 (Word 3, etc.) - released 1990
        * Office 2 (Word 4, etc.) - released 1992
        * Office 3 (Word 5, etc.) - released 1993
        * Office 4.2 (The first Power Mac-aware version; Word 6.0, etc.) - released June 2, 1994
        * Office 98 (Word 98, etc.) - released March 15, 1998
        * Office 2001 (Word 2001, etc.) - released October 11, 2000
        * Office v.X (The first Mac OS X/Aqua edition; Word X, etc.) - released November 19, 2001
        * Office 2004 (Word 2004, etc.) - released May 11, 2004

Re:GUI

[Tim Browse]

So...er, what's your point?

Re:GUI

[silverkniveshotmail.]

Major Microsoft Windows versions

                * Office 3.0 (CD-ROM version: Word 2.0c, Excel 4.0a, PowerPoint 3.0, Mail) - released August 30, 1993

Apple Macintosh versions

                * Office 1 (Word 3, etc.) - released 1990

the year 1990 came before the year 1993.

Re:GUI

[Tim Browse]

I see. So you think they rewrote all the Office apps like Excel and Word just because they started bundling them? If you want to trade wikipedia URLs, then try this...

The first version of Word for Windows was released in 1989 at a price of 500 US dollars.

The year 1989 came before the year 1990.

Also:

Word's first general release was for MS-DOS computers on May 2, 1983.

The year 1983 came quite a lot before the year 1990.

In any case, I'm not sure how any of this proves that Mac Office is a completely different codebase to Windows Office. Office is, after all, just a bundling of apps such as Word, Excel, PowerPoint, Access, etc. Just because some of the apps were released first on Windows (Word), or first on Mac (Excel), or bundled first (Mac Office), doesn't really change the fact that by Office v6, the codebases were unified. As I said, I find it hard to believe that either the Mac or Windows Office teams have re-implemented all the apps since then. Maybe I'm missing something.

Re:GUI

[silverkniveshotmail.]

I'm just saying that OFFICE came out first on the mac.
Did I miss the point? Probably. But I wasn't mistaken.

Re:GUI

[Tim Browse]

Right. Well done.

I'll bet

[Anonymous Coward]

Excel also crashes faster than Calc! *ducks*

Re:I'll bet

[Bastian227]

That's a plus for Excel then. I'd rather a program crash and get it over with than to spend three minutes crashing. Or, where you referring to the frequency of crashing? :)

Perhaps the reason is...

[mishehu]

Perhaps the reason that OO uses more private memory than does MS Office is that MS Office links to all the MS dll files, while OO bundles its own internal libraries with it?

And from article/blog/whatever: "Now to be fair, OpenOffice.org is free and is cross platform, but does this really matter to the 90% of the users in the world who only use Windows?"

If it's legally free to use and does the same task, why wouldn't 90% of the users in the world who only use Windows *not* care? People always look for what's cheaper, sometimes even if it's not better (note how MS became the company it is today...)

Re:Perhaps the reason is...

[Anonymous Coward]

It's the cross-platform part they don't care about.

Re:Perhaps the reason is...

[Anonymous Coward]

Hrmm... an extra $30 for 256MB of ram or $500+ for MS Office... which to choose which to choose...

Re:Perhaps the reason is...

[CaptainPinko]

If it's legally free to use

Most ppl don't care about that either. Buisness will but private individuals don't give too shits for the most part.

Re:Perhaps the reason is...

[elgatozorbas]

If it's legally free to use and does the same task, why wouldn't 90% of the users in the world who only use Windows *not* care?

Because apparently it doesn't do the same task. It does a much slower task. Consumers copy MS Office anyway and companies are afraid to let THE standard package go in favour of something 'dubious'...

Re:Perhaps the reason is...

[krbvroc1]

Not only that but what about all the 'hidden' services that office installs which preloads the DLLs so that office loads faster? Maybe that optimization helps? (Quicken does the same thing too).

Re:Perhaps the reason is...

[jfisherwa]

I don't think so. OpenOffice is just simply and horribly bloated.

Install MS Office in CrossOver. Run this and compare it with OpenOffice. MS Office in CrossOver is still much faster and more responsive.

Though I heard that disabling Java in OpenOffice helps things considerably?

Re:Perhaps the reason is...

[Anonymous Coward]

"If it's legally free to use and does the same task, why wouldn't 90% of the users in the world who only use Windows *not* care? People always look for what's cheaper, sometimes even if it's not better (note how MS became the company it is today...)"

Cheaper, better, whatever, it still is FUCKING SLOW. Unusable to me.

Re:Perhaps the reason is...

[Anonymous Coward]

You're talking about the prefetch feature. It's part of Windows XP, not Office. It speeds up all applications, not just Office.

http://techrepublic.com.com/5100-1035_11-5165773.h tml

Mattias

Re:Perhaps the reason is...

[ZombieRoboNinja]

"If it's legally free to use and does the same task, why wouldn't 90% of the users in the world who only use Windows *not* care? People always look for what's cheaper, sometimes even if it's not better (note how MS became the company it is today...)"

Personally, I don't think that they wouldn't not care... I think.

I doubt very many people actually pay for Office. It's either provided by their workplace or included on their PC. And most people don't seem to have a big moral issue with sharing their Office disks, illegal though it may be.

Re:Perhaps the reason is...

[Davgeary]

I use Office at work, because my employer pays for it, but at home, my machine came with WordPerfect 11 preloaded. The formatting changes when going back and forth are impossible, so I use OpenOffice at home for its compatiblility with Word. I find it adequate. Office at work seems a little faster, but I think the machine is a little faster as well. Both are pretty typical office workstation machines. I'm going to try disabling Java though...

Re:Perhaps the reason is...

[Locutus]

unfortunately, the memory is not likely to get handed to them by a neighbor, friend, or cowworker like MSOffice does. And even if it were, there's the Microsoft name, which the general population THINKs means high-tech, quality, and something they already know. And finally, most people wouldn't know what to do with the memory and would probably try to stick it into the floppy disk slot or something like that.

So, while it's cheaper to upgrade ones memory and use OO instead of PURCHASING MSO, the general population would not find it as easy as "borrowing" a copy of MSO and installing that. I've seen this happen over and over again and I wish MS would implement a annual registration system for their software because it would bump up the use of opensource software by the millions. IMO.

LoB

Re:Perhaps the reason is...

[Locutus]

they might be talking about something called osa.exe or something like it. Given that Microsoft owns and strictly controls the Windows OS and their Office software, I would not doubt that there is direct communications and involvement between those two groups. And if something does not show up in the task/server/service list does not mean it isn't getting loaded or cached because Microsoft can do this to make sure its applications run better.

IMO, it boils down to if you want one company to write ALL the software for the PC, or you would like to see many companies provide software. Microsoft can and does make sure its software has closer ties into the OS than any other 3rd party company and thus has the upperhand if all you care about is speed. But, some people are concerned that they must ask Microsoft for permission to use its software just to read their own documents. Ownership of these documents is finally becoming an issue as people realize they must ask/pay Microsoft for the right to access those documents...

LoB

Too many eyes on the code?

[thekel]

Most of the bloat I see results from kludging together work from multiple sources that are not communicating well. Can't they solve this by switching to a faster parser? Or is the format itself flawed? So many questions, this doesn't bode well. Speaking of bloat, why do linux distros come on 5 CDs with multiple versions of every possible thing. Have options is nice, but the fragmentation is getting out of head.

Re:Too many eyes on the code?

[cyclop]

The difference is: you don't have to load all applications of a Linux distro at once (haha, it would be nice to try), so you don't feel the "bloat" unless you somehow decide to open all these apps together.

Re:Too many eyes on the code?

[confusion here]

5 CDs? Try 1 CD for Ubuntu, or 60 MB for Gentoo. Distro X != All Distros.

Re:Too many eyes on the code?

[Anonymous Coward]

Can't they solve this by switching to a faster parser?

Probably, and I really hope they will do so. These tests really shows that OO needs improving. Personelly I've thought OO was a hog since it came out. I use gnumeric for my spreadsheet needs, lyx for documents and dia for diagrams. I really would like OpenOffice to be able to do the things I need - but at the moment it's just too damn slow. Gnumeric on the other hand is *excellent*.

Or is the format itself flawed?

Probably not.. I would guess on the parser.

Speaking of bloat, why do linux distros come on 5 CDs with multiple versions of every possible thing

Not multiple versions most of the time - different software! And the reason is quite simply that we - the users - demand it. I wouldn't want to use some default ftp client - I want lftp! I do NOT want to have to download it myself.

I want both firefox and opera.

I want openoffice to open those pesky word documents that simple people tend to send me. I want gnumeric for my own spreadsheeting needs. I want lyx for my documents - but I still want kword for my small notes.

Other users probably have different tastes.

This means we need standardized formats. Not standardized software.

Re:Too many eyes on the code?

[xouumalperxe]

Ubuntu is 1 CD, slackware is 2. Debian is 1 or 20+ (take your pick), and I can't recall any more right now. However, you'll find something like 4 slack CDs -- #3 & #4 are source code. I haven't used any other distros thus far, so I can't comment on any others. But I agree that the "freedom of choice" is stretched a bit thin at times (I want the freedom. I just don't want to use all of it at the same time!

When it comes to openoffice, I have to say that it is less feature rich than ms office, does indeed run slower, but on the other hand is free and came preinstalled with ubuntu when I installed it. And performance is still, by and wide, enough for me not to care (just not snappy).

shoe on other foot this time? weird.

[yagu]

It's interesting today to see the bloat and memory hog complaints leveled against the non-Microsoft product while showing MS' version as lean and mean.

I can't defend the numbers, they do look huge, but we're seeing about one or two articles a week in the trade rags about the latest memory, cpu, cache, etc. advances. Technological advances render all but the most dramatic processing demands almost moot.

In the numbers and benchmarks from this article, unfortunately, this is one of the more dramatic instances. I'm always willing to wait a little more for opening an application, or a file if other factors offset. In this case, free vs. whatever Office goes for now, typically is enough of an offset, but maybe not so for a large company where that extra "time" and computer resources add up big, and the pricing is likely to be more disounted for volume licensing.

Interesting numbers on the two different speeds on processing XML. Does anyone know or conjecture the difference in the true internal XML data for the comparison? I thought OpenOffice was the more pure in the sense that it is true human readable data in the XML while Microsoft's format is more of an envelope architecture for binary proprietary Office payloads. And, I wonder what the specifics in this test were around that.

Bottom line for me: I'm still going with OpenOffice, I've been a fan for years.

At least its not as wallet intensive!

[bazmail]

[see post title]

Bloat? Don't talk to me about bloat...

[shoppa]

1. vi is a bloated version of ex
2. EMACS stands for Eight Megabytes and Continually Swapping
3. Sometimes I just telnet to port 80 instead of using a browser

I have compiled OpenOffice from scratch... took a while!

XML

[Anonymous Coward]

sounds like an abuse of XML, which should be used for data transport and not really data storage, for that there's binary.

No Office Gripes

[afra242]

I don't use Windows and haven't since '98. At one point, I ran Linux, but kept a dual boot system with Windows, just for opening complex Word documents. Then, I started using Crossover and that saved me a lot of time and I eventually wiped Windows off my box for good.

Now I got into OS X, and I run MS Office on it. I must say though, without bias, that MS Office has to be their greatest product. It just works and I haven't ever had any issues with it at all. It is fast, user friendly, stable and usable. Let's face it: when coders code a word processor they will always look at MS Office for implementation ideas. On the Powerbook, MS Office just flies.

A few weeks ago, I tried to run Openoffice on my Debian box, and there was a huge performance decrease, when compared to running MS Office. It was certainly noticeable. It took a while for a document to open up.

Though, Office has been around for a long time and Openoffice hasn't, so I'm sure there will be lots of features and performance gains in the coming years for the latter. I'm definitely going to keep an eye on Openoffice.

Re:No Office Gripes

[qzulla]

I couldn't hang with OO on the Mac. I didn't like the X11 look and couldn't get nice fonts. I found Neoofice and love it. No affiliation. Just a happy customer.

qz

Re:No Office Gripes

[leandrod]

MS Office has to be their greatest product. It just works and I haven't ever had any issues with it

You must be a very basic user. I had plenty of users with MS Word or MS Excel files that couldn't be recovered — only option was opening an old copy, copying contents and pasting into a new document. Unless it's based on a good template, this entails lots of rework and grief. This simply doesn't happen with OpenOffice.org: the worst I've seen is needing change a troublesome font.

Re:No Office Gripes

[openfrog]

I wrote my PhD dissertation using Word back when you used two 360 KB floppies: one for the program and one where I could fit one chapter. I knew the program inside out. But then, with each upgrade it took me more and more time to stay afloat, until, after a few years, I stopped even trying to configure the damn thing to my needs altogether. They took a power user and transformed it in an utterly dependent being.

I can't understand the meaning of your assertion that Word is a good piece of software. It's unusable for me and for everyone I know.

Re:No Office Gripes

[Midnight Thunder]

I must admit that I too quite like MS-Office on the Mac, though it just has two issues for me:
    - forgetting what language the paragraph was in - I have an English system, but sometimes need to type in French
    - the embedded picture issue that affects Word files created on the Mac, but viewed on the PC - see here
The only thing that bothers me generally is closed or undocumented file formats, and Microsoft isn't the only culprit.

Re:No Office Gripes

[Benanov]

Neo Office IS OpenOffice.org + some stuff for Aqua. Same program. :)

Re:No Office Gripes

[slackmaster2000]

Parent shouldn't have been modded troll. This is a very common complaint amongst Office users, myself included. In our organization I've dealt with dozens and dozens of Word documents that just totally crapped out and would no longer open. I've also seen a whole ton of errors when tables are used in Word documents...some of them are recoverable, some of them are not. Sometimes copying and pasting into a new document works, sometimes not. Unfortunately I'm paid for this all to be my fault.

One positive thing I'll say is that Office doesn't *crash* like it used to. Back on Office 97 I think I was helping people recover documents from Word crashes quite frequently. I believe that with the release of that version, this became a much more difficult process (if memory serves me, recovering documents from temp files was pretty easy in Office 95).

Re:No Office Gripes

[syphax]

My Office gripes:

Word: Where do I start? So many quirks that can literally suck hours out of the day, especially when editing a large, complex, structured document (and yes, I do use styles correctly). Change tracking works well, except when it totally fails. I *hate* Word.
Powerpoint: Aside from the fundamental shortcomings, Powerpoint is OK, unless, say, you want more than 2 basic layouts (masters) per presentation. Or what any sort of structure beyond a basic outline. All that said, PowerPoint is better at what is does than Word.
Excel: Not too many complaints, other than it's used for lots of things it probably isn't the best tool for (e.g. statistical analysis of huge datasets). Actally a pretty good RAD tool. Overall, I like Excel a lot.
Access: Great front-end for query and report building (though formatting reports has some strange bugs), horrible, horrible, horrible built-in DB engine (that, e.g., silently truncates any data over 2GB or so, and tend to hang indefinately on moderately complex queries).
Outlook: Great for groupware features and online/offline syncing (some bugs, not terrible); terrible security and span control (even with the latest stuff).
Integration between components: Sometimes fine, sometimes horribly lame (e.g. occasional truncation of text strings after 256 characters, pasting tables between apps is rarely clean).
Scripting: Apart from the broken security model and the horrible nature of VB, the scripting capabilities of Office are pretty great.

I'm not claiming that other office suites are better, but I would say that for each task, there are generally better tools for the job. And integrating data from disparate tools isn't necessarily a big problem (I can pretty easily pipe stat. charts from R to Powerpoint, for example).

In sum, Office is not bad, but I would hardly say it is the pinnacle of what an office suite could be. We can do better, I'm just not sure that OO.o is it (I'd say it'll take 'til OO.o 5.x or so).

MS's greatest product? I'd have to vote for SQL Server. MapPoint also was an advance when it first came out.

Re:No Office Gripes

[-Grover]

I can't understand the meaning of your assertion that Word is a good piece of software. It's unusable for me and for everyone I know

      That's a bold statement. I find MS Word a wonderful piece of software for simple word processing tasks. Memos, Fax cover sheets, resumes, and regular old school work always come out formatted correctly with no weirdness from fonts, etc. Mail Merge works simply with CSV files to pop out labels and the like, and it handles internet links and embedded pictures with little to no issue. Templates even work well if you spend the time and set them up right (I have several for different things I need to write up). Is it perfect, No. Really though, what is? Utilizing different software packages for different things is part of life, but saying that Word is "unuseable" for a PhD dissertation is quite an overstatement. I'm sure there probably isn't a well defined place to cite this, but if I had to take a stab at it, I'd say the the vast majority of these types of documents are done in Word. That being said, either your "everyone I know" statement is an oversight on your part, or you simply don't know that many people. /rant -off

Re:No Office Gripes

[99BottlesOfBeerInMyF]

I must say though, without bias, that MS Office has to be their greatest product. It just works and I haven't ever had any issues with it at all.

I've used Word casually and professionally on both Windows and Mac OS X. I also use OpenOffice upon occasion. Without trying to compare the two, here are few problems I've had with Word that have prompted me to avoid it:

• Closed, ever changing file format - I have old word files I inherited that don't open in any program I can get my hands on, and others that only open in OpenOffice. These were created on an old version of Word. I don't trust .doc files to be usable in the future.
• Large documents - at about 200 pages with the occasional graphic, Word consistently fails to properly save or open files. Sometimes it will save a file and the file will no longer open at all. Sometimes it will corrupt the document beyond all hope of recovery. At one point I was saving files, closing them, and then re-opening them before making a back-up every time I edited the files.
• Output formats - In order to get decent XML, properly formatted from word you need to buy an expensive add on program, like Webworks pro. Ditto for usable HTML.
• Formats stored in carriage returns - what could be more annoying than storing all the formatting info for a paragraph in the carriage return of that paragraph? Why does all this info disappear if I merge two paragraphs. How can this have not been fixed yet?
• Images - Word messes up images for me regularly and inserting large numbers of images consumes enormous resources, making the whole program slow down considerably.
• Spellchecking, grammar checking, translation, dictionary, thesaurus - Word still can't use the standard services available to other applications on my system forcing me to use their inferior spellchecking, etc. or to copy and paste text out of word, into another program, and then back if I want to lookup a word in an online dictionary, or translate a paragraph from or to another language.

All of these are reasons why I can't use Word as a professional solution for general text/document editing. That is not to say that OpenOffice is any better, it has plenty of its own problems, but if you haven't run into any of these using Word then you must not use it very much or for a large variety of tasks.

Re:No Office Gripes

[xtracto]

I think what you want to say is that Office Suites suck overall.

I agree with that, I use Latex (Tecnixcenter) to typeset documents, I use gnuplot or Graphcalc to create graphics and mysql with java for databases (I dont know any scripting language like python or tcl/tk... I will learn them one day...).

Basically, what does an Office suites provides:
- Writing (Tex... or Abiword if you like WYSIWYG)
- Statistical oriented Data management (you could use R)
- Database oriented data management (Use mysql, or any other DB management, even Access!!)
- Mail (I use only webmail [gmail] but feel free to use anything)

The fact that with the Writing subapp of these office suites you can do all 4 is incredibly bad.

I remember that, once, there was an opertaing system and a community whoes trend on applications was to write simple, stand alone, task oriented applications whose results could be combined to make something big.

I am sure that is possible to do making use of Graphical User Interfaces!
And, I am also sure that if the approach to program was that, applications will tend to be a hell of more stable.

The only downside I see on that is that there should be a need of a lot of standarization in the different output/input formats. But I think this is not difficult now with XML.

Re:No Office Gripes

[dominator]

Though, Office has been around for a long time and Openoffice hasn't, so I'm sure there will be lots of features and performance gains in the coming years for the latter. I'm definitely going to keep an eye on Openoffice.

That's not true at all. While OpenOffice is "only" maybe 5-6 years old now, it is built on top of the older StarOffice codebase, which has been in development since the mid-1980s. It's not like they started from scratch a few weeks ago...

http://en.wikipedia.org/wiki/StarOffice

Re:No Office Gripes

[99BottlesOfBeerInMyF]

Oh wait, I forgot one. Spanish menus. I used to run Word in English, but every now and again the spell checker would run across an acronym (I think IS-IS was one of them) where it would suggest it was an incorrectly spelled word in the Spanish language. At this point all the menus would convert to Spanish and resetting the language to some other language would fail with an error. The only fix was to quit the program and restart it. That one drove me nuts!

Re:No Office Gripes

[Anonymous Coward]

MS Office has many problems - and what's disgraceful about them is that most have existed for more than 10 years. Here are a few of my worst offenders:

Word: (1) when the hell are they ever going to get flowing, anchoring, and captioning to work correctly? Try to anchor a picture on a page, add a caption beneath that figure and then flow the text around top & bottom. It still can't be done - the best solution is to add a text box but then your figure number don't update properly. This is so utterly basic and inexcusable that they haven't fixed this.

(2) What about cross referencing equations? There appears to be a facility for this but damned if it works - you have to enter friggin' bookmarks to cross reference equations.

(3) What's up with page layout view? Does it actually print the way it looks in page layout - NO! Can't they get that right? Totally unacceptable given the amount of time this product has been on the market. I would be ashamed to say I wrote this software.

There's more but I'm busy - I could go on about headings, etc.

Excel: sorry to disagree but charting in excel simply blows. I can't tell you the number of times I've wanted to chart along three dimensions - can't do it in excel. Also, copying charts from excel to word - has bugs - if your line types are dashed they got copied as solids - that's a nice feature. There's also a lack of support for text subscripts and superscripts when labeling charts - not good for scientific use. Again there's more but you get the point.

Powerpoint: Drawing in PP sucks. After you meticulously align objects on slide A and then move to other slides and return to slide A, somehow lines have decided to move on their own? This happens even after saving the document - don't know why, it just does.

There is plenty of opportunity to improve upon MS Office - if you use it enough you'll realize that. Several other products are almost there and I can't wait until they surpass this dinosaur crapware that most of the world accepts.

Re:No Office Gripes

[mrchaotica]

He's talking about Office/Mac. Even though I'm biased against Microsoft, I can admit that the Mac version is pretty good (better than the Windows version!).

Re:No Office Gripes

[Anonymous Coward]

MS Office has to be their greatest product

I think they topped out with Notepad. ;-)

(ducks)

Re:No Office Gripes

[mindstrm]

Rumor has it that the team that did Office for OS X actually caused a bit of a stir because it was better than Office for Windows.

Re:No Office Gripes

[ywwg]

except for the bug we found, where if you leave Word in Office X open for 6 hours or more its internal state becomes corrupted and you can't save or print the document. So you have to kill the program and start from where it last saved correctly.

Re:No Office Gripes

[Alcilbiades]

The biggest difference I have noticed between OO and MS O is file size. Just for kicks save the same file in both office bundles and see which one is bigger. The MS file is typically 4x larger to store than the OO one.

Re:No Office Gripes

[ByeLaw]

I am a Network/Systems Engineer for a large college and I use OfficeXP in many of our presentation and data distribution solutions for our MIS. I have also nearly 10 years of experience with this product and automate many of the functionality with other programs to a high degree. I admit, early versions where buggy but after Office 97 things went more smoothly and now I don't have any real problems.

As a network Engineer, I have not seen any users with problems as you suggest and we do have expert users so we are not just using the basic features of the software. Maybe you have an axe to grind?

Re:No Office Gripes

[prisoner-of-enigma]

Closed, ever changing file format - I have old word files I inherited that don't open in any program I can get my hands on, and others that only open in OpenOffice. These were created on an old version of Word. I don't trust .doc files to be usable in the future.

I find this the most interesting demerit you assign to Word because Microsoft hasn't "broken" it's file format since it moved everyone from Office 97 to Office 2000. The current Word 2003 is completely backwards compatible with all prior Word document formats. Perhaps you failed to load the appropriate import filter when you first installed Office? Otherwise I cannot fathom why you can't open these files. We've been using Office since well before Office 95 and are now on Office 2003 Pro. There are ten-year-old documents out there that we have no difficulty opening whatsoever.

Large documents - at about 200 pages with the occasional graphic, Word consistently fails to properly save or open files. Sometimes it will save a file and the file will no longer open at all. Sometimes it will corrupt the document beyond all hope of recovery. At one point I was saving files, closing them, and then re-opening them before making a back-up every time I edited the files.

This could potentially be a machine resource issue and not a Word issue at all. We have purchasing contracts that run in excess of 200 pages and we've never had any difficulty opening such files. Our machines are Dell Precision 370 workstations with 3GHz P4's and 1GB of RAM, but even older systems with a 733MHz P-III and 512MB of RAM don't have any problems.

Output formats - In order to get decent XML, properly formatted from word you need to buy an expensive add on program, like Webworks pro. Ditto for usable HTML.

You'd have to quantify what constitutes "decent XML" and "usable HTML" for this comment to make much sense. Word 2003 can export in both formats and they look just fine to us.

Formats stored in carriage returns - what could be more annoying than storing all the formatting info for a paragraph in the carriage return of that paragraph? Why does all this info disappear if I merge two paragraphs. How can this have not been fixed yet?

I haven't experienced this issue, so I can't comment on it. However, if it's true and not peculiar to your installation I could certainly see why it would be annoying.

mages - Word messes up images for me regularly and inserting large numbers of images consumes enormous resources, making the whole program slow down considerably.

The nebulous "messes up images for me" comment leaves much to be desired. We place bitmaps of all sorts in Word docs with no problem at all. Admittedly it does make the file size bloat dramatically, but that's hardly Word's fault as graphical data is obviously much larger than textual data. The slowdown could again be a machine resources issue since Word may no longer be able to hold the entire contents of the document in RAM at once. Our responses are snappy even with lots of images, although it is slightly slower than text-only.

Spellchecking, grammar checking, translation, dictionary, thesaurus - Word still can't use the standard services available to other applications on my system forcing me to use their inferior spellchecking, etc. or to copy and paste text out of word, into another program, and then back if I want to lookup a word in an online dictionary, or translate a paragraph from or to another language.

I'm a bit confused as to what constitutes "inferior spellchecking." Are you saying the spell check feature does not work? Obviously that is not true. Are you saying it doesn't do what you need it to do? Well, that's very subjective so you can hardly call it a generalized problem with Word, now, can you? If you don't like the default dictionary, you can customize it.

As for "standard services" elsewhere on your system, I believe it's up to the third-party providers to figure out how to integrate their software into Word, not the other way around.

Re:No Office Gripes

[99BottlesOfBeerInMyF]

The current Word 2003 is completely backwards compatible with all prior Word document formats. Perhaps you failed to load the appropriate import filter when you first installed Office?

Yes perhaps I, and everyone else running on multiple platforms has incorrectly installed Word. Mind you, some of these documents do open in OpenOffice and in very old versions of Word (that are no longer available for purchase). As near as I can tell most of them were created on Office 95 and 97 for Windows, and Office 2000 for the mac.

There are ten-year-old documents out there that we have no difficulty opening whatsoever.

I suspect your documents do not contains as many interesting uses of tables and other less used features. In any case I'm certainly not the only one with these problems, you can find them discussed on many forums. I know I found plenty of problems and few solutions last time I looked.

This could potentially be a machine resource issue and not a Word issue at all. We have purchasing contracts that run in excess of 200 pages and we've never had any difficulty opening such files.

I doubt it. Machine specific problems do not usually show up on such a wide variety of desktops and laptops, nor are they usually cross-platform. Perhaps you do not have as many graphics embedded in your files? I think we ran into problems at about 200 pages with one graphic every 10 pages or so.

You'd have to quantify what constitutes "decent XML" and "usable HTML" for this comment to make much sense. Word 2003 can export in both formats and they look just fine to us.

Are you joking? The HTML coming out of Word looks like it was written by a deranged Klingon. It is unreadable and unmaintainable. As to the XML, XML with large chunks of arbitrary binaries embedded are not very useful. XML tags in a specified format with specified attributes, as based upon the paragraph and character styles are what would be useful for a real documentation project. There is a reason companies pay Quadralay a thousand dollars a seat to get decent XML from Word, that reason is that Word's XML is useless.

I haven't experienced this issue, so I can't comment on it. However, if it's true and not peculiar to your installation I could certainly see why it would be annoying.

This has always been true, how can you not have noticed it? The only people I know of who are not aware of this are people who have only used Word, and thus except it as "normal" for a word processor.

The nebulous "messes up images for me" comment leaves much to be desired. We place bitmaps of all sorts in Word docs with no problem at all. Admittedly it does make the file size bloat dramatically, but that's hardly Word's fault as graphical data is obviously much larger than textual data.

First, placing images can be done by linking or importing and it makes a big difference, as does the file format. It is annoying to have to reformat all graphics in multiple formats, for the web and for print because the editor cannot handle them properly. Second, a reasonable layout application allows you to view lower quality versions of the images you embed so that it does not suck up all the RAM caching them and thus bring the system to a slow crawl, as it approaches the available system RAM limit.

I'm a bit confused as to what constitutes "inferior spellchecking." Are you saying the spell check feature does not work? Obviously that is not true. Are you saying it doesn't do what you need it to do? Well, that's very subjective so you can hardly call it a generalized problem with Word, now, can you? If you don't like the default dictionary, you can customize it.

Yup, then I can have one customized spell checking library for Word, and one customized spell checking library for every other application I use. Talk about useless duplication of effort.

As for "standard services" elsewhere on your system, I believe it's up to the third-party provide

Re:No Office Gripes

[leandrod]

after Office 97 things went more smoothly and now I don't have any real problems.

In my experience, users tend to route around known bugs, for example advanced automated formatting in MS Excel or complex layouts with styles in MS Word.

I had huge problems until MS Office XP SP1 at least. SP2 helped a lot, but I haven't used complex documents since then.

Re:No Office Gripes

[00110011]

Or maybe they have a lot of machines with bad memory sticks.

Re:No Office Gripes

[leandrod]

He's talking about Office/Mac. Even though I'm biased against Microsoft, I can admit that the Mac version is pretty good (better than the Windows version!).

True enough, just not as reliable as OpenOffice.org. At least with pre-Mac OS X versions of MS Office I still lost some complex document formatting in MS Word.

Re:No Office Gripes

[NutscrapeSucks]

At least in my experience the Mac versions of Office is not nearly as fast or stable as the Windows versions, although it has a few unique ease-of-use features. It also has its own set of slightly-incompatible file format issues.

On its face

[shareme]

On its face MS Office does not handle OpenDocument format so theother claism are entirely suspect.. Do anly /. contribs actually read what they submit?

Re:On its face

[Anonymous Coward]

If OpenOffice is performing reasonably for the OpenDocument format, then OpenDocument must be technically inferior to Microsoft's own formats.

So: either OpenOffice is bloated and slow, or Microsoft is right to not support a technically inferior document format like OpenDocument.

You choose.

Re:On its face

[Anonymous Coward]

Uh, they did it with an excel file vs. the same data in an OO.o file. Did you read it?

My results

[travail_jgd]

I get the exact opposite results for Excel and OpenOffice Calc: Excel takes forever to load, doesn't share memory, etc, and Calc is a lot faster/leaner.

Then again, I'm running Excel in Crossover Office; all those Windows libraries aren't "preloaded" for me. Maybe that's why XP and Vista have such large system requirements?

Re:My results

[GiMP]

> Then again, I'm running Excel in Crossover Office; all those Windows libraries
> aren't "preloaded" for me. Maybe that's why XP and Vista have such large system
> requirements?

Windows does do something to this end. I forget what they call it, but the OS will automatically determine which applications are used most frequently and will do some sort of speed optimizations for them. I'm guessing it either prelinks the applications or keeps some dlls in memory.

On another side of this.. I wonder how many people complaining about the speed of OpenOffice (on Linux) have attempted to prelink it first? Some claim it can increase the speed by 50%.

It appears that Red Hat Fedora uses prelink by default.

Re:My results

[Anonymous Coward]

How big a file you are using to test? I used a spreadsheet that has 10 pages, 30000 cells each, 9 out of 10 pages are functions. I changed a single numbers, it took 30 sec for the system to response, and it does not always refresh poperly. What spreadsheet program am I using?

Re:My results

[Anonymous Coward]

You're running Excel *and* Crossover Office, *and* Excel is not a native app, and you're comparing it a native app. And you think it's informative that the native app is faster?

Re:My results

[FidelCatsro]

I don't really use Windows so I may be wrong here . I think it has something to do with the prefetch folder in C:/windows . The OS takes some form of snapshot of the commonly used apps and loads it and some of their DLLs at start-up

Re:My results

[zooblethorpe]

You're running Excel *and* Crossover Office, *and* Excel is not a native app, and you're comparing it a native app.

To stretch the point a bit, isn't this essentially the same complaint against TFA, just the other way round?

Re:My results

[Anonymous Coward]

No - Windows is an officially supported platform for OpenOffice. OpenOffice for Windows is compiled on and for Windows. OpenOffice for Windows does not rely on a complete API reimplementation on top of an unsupported OS.

Re:My results

[slavemowgli]

Oh? I wouldn't be so sure about that. Sure, windows is not unsupported, but don't you think that OO.o is likely to implement some sort of API abstraction layer on top of the OS to facilitate porting to different platforms? That's certainly something that can cause bigger memory and CPU time requirements, and it's also something that MS Office obviously does not need (yes, there is also MS Office for the Mac, but IIRC, they're actually relatively independent products with different codebases that just share the same name for marketing reasons).

Re:My results

[dave420]

XP? 128MBs is needed. That's all. It doesn't have massive system requirements. Sorry to pick you up on that, but I feel bullshit on all sides is still bullshit.

PS I just fired up excel, and it takes 3 seconds to load. And I use office software very rarely, so nothing was pre-loaded (and I clear my precache frequently, so it wasn't in there ;)).

Re:My results

[travail_jgd]

"Needed" and "optimal" are two different things.

I don't have a lot of hands on experience with XP, but Win2K was clunky with 256 MB (on an Athlon 1600+). Once the RAM was upped to 512 MB it seemed to be happier, becoming much more responsive. That's what I'm basing my statements on.

The "minimum requirement" is a Public Relation numbers game, anyway. Too high and potential customers may be alienated, too low and the poor performance could fall into the 'legal action' category. I have little doubt that XP could be booted with 64 or 80 MB of RAM; whether or not the system would be useful is a different story.

Re:My results

[sparkz]

For the first time in many, many years, I've got to use Windows and MS Office.
I used to have a 256Mb 1.8Mhz Intel P4M with a 5400rpm HDD on which I ran various Linux distros, most recently Debian for the past year or so. I was using OOo to edit StarOffice Writer documents. It took maybe 2-3 seconds to load a relatively complex 30-50 page document.
I've now got a 1GB 1.6GHz Intel P4 Centrino running Windows XP, and I work with MS Office documents. I find that initial load is pretty quick - the CPU is a bit slower, but it takes a similar amount of time to open a document - given certain conditions:

• If the PC has only just booted up, I can log in and quickly see the desktop. However, any clicks in the first 10 seconds or so are apparently ignored. It looks like a PC ready to use, but it isn't.
• If I have more than about 8 windows open (typically MS Outlook, MS Word (maybe 4 docs), MS Excel (1 or 2 docs), Jabber (1 or 2 chats)), then clicking on a .doc file takes a good 10 seconds to load. In the meantime, there is no indication whatsoever that the system has acknowledged my click and is making any attempt to open the document (at least OOo has an activity bar showing that it is processing the document) - not even the disk-activity indicator light on the laptop, so it's not that the HDD is slow (not sure of the speed, but can't be worse than 5400rpm), it's just sitting there, thinking about it.
• Half the MS Office docs I view automatically show revision changes - I don't care, and there doesn't even seem to be a keyboard shortcut to kill it. I have to go to a menu or toolbar to view the fscking document itself.
• Having stripped an example document of its detail, mailed the stripped-down version to someone else, and received his updated version, I can't open that document without it constantly asking me if I want to merge his changes back into the original document ... even after I delete the original, it still asks me. No, I don't want to merge it. It's a new document. That was a fscking template. Leave me alone.
• This may be more indicative of corporate documents, but most docs I receive ask me to enable macros, update external links, etc - I'm not stupid enough to automatically enable such things, but even in docs where I know they are not necessary, I still get asked. That's another meaningless question (no option to view the macros before enabling them, for example) which gets in the way between myself and the document.
• Document views - this seems to be defined as part of the document itself, whether it's 100%, 75%, page-width, or 2-pages-per-screen, or even 4 pages per screen. Totally unintuitive. I'm sure there's a way to get the view you want, but it took me long enough to work out that "hover the mouse over the (alleged) pagebreak and click on "show whitespace" to view each page as a page.

So that's my "I've had 1 month with MS Office after 6 years with StarOffice" summary - most reviews seem to be the other way around. I'm sure I've missed out most of my favourite rants, but that's just a sample of the frustrations I've had so far. Bring back StarOffice - it's been working great for me for the past six years.

Re:My results

[NutscrapeSucks]

Well since OOo doesn't even run on the Mac, maybe they should get rid of this huge bloatastic "API abstraction layer" they have burdened themselves with and just target the native Windows and X11/GTK APIs.

"It runs on some other platform that you don't use" is not an excuse that flies with end users. MS learned that the hard way with MacWord 6.

Re:My results

[ccp]

I wonder how many people complaining about the speed of OpenOffice (on Linux) have attempted to prelink it first? Some claim it can increase the speed by 50%.

Why, oh why?

Puzzled by so many posts about OO speed I tried it right now:

OOWriter 2.0 from clicking icon to page open with cursor blinking, 7 seconds.
In a very middle of the road machine: Atlon XP 2400, 768 RAM.

Why would anybody need an even faster start?
What practical difference would it make?

OOs slowness is some kind of urban legend, kept alive by MSFT astroturfers.

Cheers,

I'm unsure where the truth lies...

[erroneus]

...but I am sure that if there is any truth to this, it will be addressed and corrected in a fairly short time. I believe it's really as simple as that.

I think it would be really nice of the people producing the article to make the sample data available to the rest of us to see the results for ourselves.

Re:I'm unsure where the truth lies...

[erroneus]

...oops, I just realized the files are linked in the article... the link word was a document file type and I assumed it was a link explaining the type of document... not a link to the documents themselves.

Re:I'm unsure where the truth lies...

[networkBoy]

"I think it would be really nice of the people producing the article to make the sample data available to the rest of us to see the results for ourselves."

Surely you RTFA didn't you?
They did make the data available dufus.
-nB

not fooling anybody

[Anonymous Coward]

I'm not surprised, I heard MS put secret anti open office code in their latest security updates to fool the sheep into using their stuff because it 'appears' superior in every way. To that I say HA! You've not fooled me!

Call a Spade a Spade

[espek]

Just go ahead and admit it, they both suck for different reasons. We need a third player.

Re:Call a Spade a Spade

[IWantMoreSpamPlease]

AbiWord
Lotus SmartSuite
GoBe Productive

Re:Call a Spade a Spade

[digidave]

KOffice. Not as cross-platform as OOo, but very good and very fast.

Re:Call a Spade a Spade

[tedgyz]

Just go ahead and admit it, they both suck for different reasons. We need a third player.

Now THAT is the smartest thing I've read on this topic yet!

I've tried to use OpenOffice, but it seems that I've traded one feature-bloated app for another. Apart from being free, I have not found any upside to OpenOffice. But... Since Microsoft seriously thinks that I should pay ~$300 per copy of Office, it is compelling. Given that my family now has 5 PeeCees, I don't think MS will be getting my $1500 any time soon.

As far as performance... OpenOffice suxx0rs. Even with the "quickstart" running, it takes forever for Calc to startup with a blank sheet.

Now on the the third choice... Anybody? Please? Let's call it SmallAndFastOffice to be sure we get it right this time.

Re:Call a Spade a Spade

[Anonymous Coward]

WordPerfect? Yes, it's still around. It loads quickly. You can decide what parts of the program you don't want to install. And reveal codes, reveal codes, though for all I know Open Office can do that too.

Re:Call a Spade a Spade

[Bastian]

AbiWord: Not an office suite! (Though a much better start for a word processor than OO.o writer.)
Lotus SmartSuite: Windows-only.
Gobe Productive: Windows-only.

Re:Call a Spade a Spade

[Anonymous Coward]

Gobe Productive is Windows only? I know that they ported it to Windows but had no idea people had forgotten that it started (AFAIK) on BeOS, and is still bundled with Zeta. I always thought the name GoBe was indicative of that. Their website even says they've started a Linux port as well.

Re:Call a Spade a Spade

[Anonymous Coward]

GoBe Productive is a very nice app. However, the last milestone reached on their frontpage was almost 4 years ago. I have a feeling that the product is essentially dead.

Re:Call a Spade a Spade

[Bastian]

They started that Linux port years ago. I think the chances of them finishing it before they run out of money are slim to none at this point. The BeOS version is no longer being sold, and has not been sold for years.

Whatever it was, whatever it may be, right here and right now, Gobe Productive is for Windows only.

Re:Call a Spade a Spade

[linguae]

Well, AbiWord and GNUmeric are part of a collection of applications known as GNOME Office. They use GTK2, and are very lightweight (even though they don't have all of the features and capabilities of OpenOffice; AbiWord needs some work on its Word import/export functions). A GTK2 presentation application would be a nice thing.

On KDE, you have KOffice, which is pretty lightweight (compared to OpenOffice) and has a decent mix of features.

Now if somebody made an FOSS office suite for GNUstep and OS X....

Re:Call a Spade a Spade

[misleb]

Just what we need, a third application suite that has 10x more functionality than most people will ever use.

-matthew

Re:Call a Spade a Spade

[Hugonz]

It's called KOffice...and with Open Document... this may be the time for app independent document processing.

Re:Call a Spade a Spade

[Kynde]

Just go ahead and admit it, they both suck for different reasons. We need a third player.

Patience young padawan. So far the biggest problems with OO have been the lacking features compared to the M$ Office aswell as interoperability with the M$ Office. We're obviously getting somewhere now that people start benchmarking and complaining about memory usage. Seriously, five years ago no one would've even bothered to check memory usage when comparing those products, there wasn't much to compare.

For the record, I'm not saying OO ain't bloated, so it seems, and perhaps there's been too much pressure to reach interoperability and feature richness, but it's too early to condem it. Time will tell wether their internal design is good or not. Can it be made faster/leaner/meaner without too much sweat and tears...

Re:Call a Spade a Spade

[petabyte]

And though everyone is always hung up on OpenOffice, I made it through college with Abiword. Today I use Abiword and Gnumeric pretty much exclusively as my office apps. Ocassionally I'd like to have something that lets me do powerpoint (gtk app ideally). With evince supposed to be getting the support to view PPTs though, thats one step closer.

Also, abiword runs on OS X and I know work is being done to pretty much put gnumeric everywhere. That'll be a good day :).

Re:Call a Spade a Spade

[Arandir]

Bloat and speed are problems, but are they the huge problems everyone here is making them out to be? Startup time is abysmal, but once it's started, opening a document never takes more than two or three seconds for me. Responsiveness is never a problem.

Yes, it falls behind MSOffice in this regard, but it makes up for it in spades. It's major advantage is that it's EASIER to use! Those weaned and raised on Word may not see it, because to them there is only One True way to write a document, but OO.org does styles and templates right. It lets you edit your content without having to spend half an hour twiddling with the layout. It doesn't get in your way!

Re:Call a Spade a Spade

[Henk Postma]

I for one would love to see Rasterman take a look at the OpenOffice code after he's done with Enlightenment 17. Running e17 now and it is 1) the prettiest 2) the fastest DE I have tried.

I can only imagine how nice it would work if he could use the evas framework to render presentations and other screencontent.

Re:Call a Spade a Spade

[dtfinch]

I was turned off to it a few years ago, because it was ugly, buggy, and slow in my opinion, but now it seems KDE has quickly grown into one of the greatest things since sliced bread.

Re:Call a Spade a Spade

[blibbler]

Hmm, forgive me if I misunderstand you, but are you saying that Open office is both bloated, and feature-lacking? That doesn't sound like much of an endorsement to me.

Anyway (according to wikipedia), Open Office has been around for almost 20 years as Star Office (although it was only integrated about 11 years ago), while Microsoft Office was first released as an "integrated" package about 12 years ago. It is misleading to suggest that Open office hasn't had the time to compete.

Re:Call a Spade a Spade

[nortcele]

Just go ahead and admit it, they both suck for different reasons. We need a third player.

Agreed. But we should just leave US politics out of this discussion. We'll never get a viable third party.

Re:Call a Spade a Spade

[Anonymous Coward]

Only if you believe what you read on slashdot. Warning: reality may differ significantly.

Re:Call a Spade a Spade

[Anonymous Coward]

Only being sold for Windows? Ten seconds on Froogle says otherwise. I'd even guess that more copies of Zeta (which include GoBe Productive) are sold than copies of the Windows port. I also expect that more people still use the BeOS version than the Windows version. The BeOS market might not be viable, but that doesn't mean people don't still use it. I know that the Linux port was never completed, and only mentioned it since I assume you mean "not for Linux" when you say "Windows-only". BeOS isn't dead despite what most people think. It's undergoing active commercial (YellowTab) and open source (Haiku) development and still has a rather loyal group of users that I'd guess is increasing (few people quit using it entirely, and, given the forum posts, several people are trying it for the first time). I've never used the Windows version of GoBe productive, but I have used GoBe Productive for BeOS just a couple months ago (my printer doesn't work in BeOS so I don't use it for schoolwork).

FUD

[Moby Cock]

This is total garbage. I have been using OO 2.0 at home since it was released and I have noticed no lag compared to Office which I have to use at work. I do not have quantitative numbers to present, but I can certainly attest anecdotally that this blogger is flat-out wrong. I notice no appreciable speed difference between the two suites while processing the same files. (The machines are roughly equivalent at home and office)

Re:FUD

[bogaboga]

To me, OpenOffice.org is fine except for its beauty. It's just plain ugly on Linux systems! How can one make it a pleasure to look at just like SUN's StarOffice or OpenOffice are on Windows systems?

NeoOfficeJ

[ontheheap]

I recently purchased an iBook G4 which came with a trial edition of Office.Mac (or whatever it's called). I used it for the 45 days of the trial and then switched to "OpenOffice.org for the Mac," otherwise known as NeoOfficeJ. The only thing I've noticed thusfar is that Neo takes about 1.5 times longer to run initially, and it seems to take longer to save files. Other than that I really haven't noticed any other differences in performance.

Re:NeoOfficeJ

[Anonymous Coward]

Got NeoOffice/J installed on a Mac here and it seems pretty sluggish. Some of the dialogs are also pretty inconsistant with the rest of the OS, though I expect both of these issues to be taken care of as the product matures. Pretty damn good right now, though.

Re:NeoOfficeJ

[plj]

As OpenOffice 1.1.x (and thus NeoOffice too) handles fonts internally, it has to load native fonts during startup. This can take a long time if you have lot of fonts. OO.o/X11 should start faster, but it may has other problems. See this thread for more information.

So why isn't it more popular?

[Anita Coney]

Americans LOVE bloat. All you-can-eat restaurants. Trucks with HEMIs and lift-kits only driven on pavement. Giant houses with even larger pole barns. You'd think it'd be a little more popular.

Re:So why isn't it more popular?

[Anonymous Coward]

You know what they say about guys with giant trucks, pole-barns and Office suites...

Re:So why isn't it more popular?

[Jestrzcap]

Fast Food? Fast cars/trucks? Ummmmm fast..... garage doors?

We liked bloated speed I think.

Re:So why isn't it more popular?

[Anita Coney]

"Bloated Speed" Sounds like a late 80s speed metal band featuring four blonde fat chicks.

Re:So why isn't it more popular?

[Anonymous Coward]

Loans, mortgages and still can't write?

parsing a large XML in java can be a hog

[n0md]

Yes, potentially if they are parsing the XML into a full DOM tree, this could potentially take MASSIVE ammounts of memory. If you read up on the DOM parser for java. There are clearly many benifits of parsing an xml document into a DOM tree structurally and whatnot, but practically, at this point, if your document is large its just not viable memory wise. I don't know if this is what OO does, but its just a possible explaination.

MOD PARENT OVERRATED

[Anonymous Coward]

Hypocrite - show some repeatable results, don't just talk out of your ass!

No Methodology

[anderm7]

These articles are complete garbage. No mention of methodology is made. What files were loaded, what conditions were they loaded under. Was it the same machine, or a very similar machine. What distro, what JVM, and on, and on, and on. Sounds like another MS shill to me.

Re:No Methodology

[archen]

I wouldn't say they're garbage just no methodology for pointing out the obvious. MS Office 2k spanks OO.org on Windows on every machine I've tried it on - on both speed and memory.

Besides which, if there are that many vairables to OO running "well" then at least you could say MS office is consistent.

It doesn't really matter to me since I'll be using OO anyway. Besides which now that the open source world (Koffice, etc) have also pleged to support the OASIS format, we should be able to pick and choose our word processors in a few years without worrying about compatability. Open office isn't our last hope, it's our foot in the door.

Re:No Methodology

[Bastian]

Yeah, because, you know, nobody else has been complaining about how slow and bloated OO.o is, or how slow and bloated StarOffice was. OO.o is perfect. Everyone in the Linux community is just perfectly satisified with their choice of office software. I know I am.

Re:No Methodology

[RandomPrecision]

I have to agree, particularly since OpenOffice loads faster on my computer than the MS Office 2003 trial, or whatever it is. And it doesn't go away after x days.

I guess some people seem to like MS Office better, but I think being able to type a formatted paper that can one can print is something that all computers should have for free.

All of my computers came with MS Word for years, and I never thought of the possibility that they could charge for something like that, until just a few years ago, PC's started coming with only a trial version of Word. And that's why I've switched to OpenOffice - I don't think that every family with children that need to type reports for school suddenly owed Microsoft a few hundred more dollars for an Office Suite.

Re:No Methodology

[diegocgteleline.es]

What files were loaded, what conditions were they loaded ..except that they provide you links to them (did you _read_ the article)

Was it the same machine

Yes

What distro

No distro - it was openoffice for windows

Sounds like another MS shill to me.

And you sound like a FOSS shill to me.. ;)

Openoffice is bloated. Anyone who has used it should realize. We know it, but we have no OSS alternative to it, so...

Re:No Methodology

[Anonymous Coward]

Slashdot needs a +1: Sarcastic mod.

(And a -1: Sarcastic mod)

Re:No Methodology

[Anonymous Coward]

Slashdot needs a +1: Sarcastic mod.

(And a -1: Sarcastic mod)

Yeah, I can't think of anything we need more.

Bought

[Psionicist]

He is already anti-Open Document http://government.zdnet.com/?p=1723 and heavly pro-Microsoft http://blogs.zdnet.com/Ou/ so this is not unexpected.

Re:Bought

[Anonymous Coward]

Thats what I thought at first, but I downloaded the test data - and its true. I have Office XP (not 2003) and OO 2 and MS is WAY faster... I tend not to notice since I usually work on a much beefier machine, but here it is certainly a big deal.

Re:Bought

[Anonymous Coward]

What's not unexpected?

If it's the truth, then it's the truth.

I certainly would not take what any PRO OO.o huckster said with any more confidence.

You think Sun or OO.o would point out MASSIVE shortcomings in their own products?

Or would you rather a world where no one expresses an opinion different from your own no matter what?

I'll tell you what's not unexpeced, to have 90% of /. readers completely ignore facts because the source is "one of THEM".

Re:Bought

[Anonymous Coward]

Why did you get modded Insightful? I should just link to http://www.yahoo.com/ and make some baseless claim that supports the /. groupthink and get modded insightful too. Apparently no one bothered (including you) to really read the link you provided.

How is this guy a Microsoft shill? (emphasis mine)

From your linked article:

George Ou thinks Massachusetts is in a no-win situation, in requiring government applications to support OpenDoc.

Mandating the adoption of the Oasis Open Document format is kind of like mandating the conversion of all public documents from English to Esperanto. Even though English was a second language for me, I'm perfectly happy with it being the de facto standard and I suspect that most people are in no mood to learn another language because some State bureaucrat mandates it.

On the other hand, he points out:

Since Microsoft is already opening up its Office 12 XML formats (with some restrictions) and the existing Microsoft formats are already semi-open, why not officially open up their existing file formats and be a hero. Most people don't stick with Microsoft Office because they're locked in to the file format, they use it because that is the tool they learned and it works well. There is nothing to be gained by anyone from a file format war.

So while you may not agree with his conclusions, he hardly sounds like a paid shill to me.

Re:Bought

[d99-sbr]

So you're saying the memory and CPU usage of OO.o is dependent on the user's opinions?

Microsoft Office really blows the competition right out of the water IMHO. Together with the awesome SQL Server they're easily the best products out of Redmond.

Re:Bought

[Tony]

Microsoft Office really blows the competition right out of the water IMHO. Together with the awesome SQL Server they're easily the best products out of Redmond.

I believe that's called, "Damning with faint praise."

Re:Bought

[Anonymous Coward]

Informative not Insightful... must... use... preview... button...

My point is still valid however.

b*tch, b*tch, b*tch

[dougwhitehead]

As with any technology, first it becomes capable, then it becomes efficient.

The fact that people are carping about efficiency means that it largely has what they need.

There is a fax available!

[Luscious868]

There is a fix available here if you run Windows:

http://www.microsoft.com/office/

For those running Linux, you'll also need this:

http://www.codeweavers.com/products/

Re:There is a fax available!

[m50d]

Lol, have you ever used COO? OOo is slow, yes, but I think even it would be defeated by that.

is the solution "Mozillarization "?

[ghee22]

read about it here: http://www.oreillynet.com/pub/wlg/8136.

I'm not sure. I know I rarely use anything except writer, so maybe having a writer lite edition as well as the whole suite.

Measuring

[pyta]

There is question what exactly these results mean. Pure memory usage is very hard to compare because Microsoft Office is based on their own API and COM technology whereas OpenOffice use UNO. It means that whole UNO runtime is loaded into memory and is part of the soffice process. The same problem is in startup times - running Openoffice means to run almost whole OS. Of course that users don't have to take care about such argumentation, but I think that we cannot expect any major decrease of memory usage in any near future.

Re:Measuring

[SPY_jmr1]

So....

You're saying that if OO.o is bloated, and that if it's using UNO, it is basically pulling too many Draw Four cards when it's turn comes?

Easter eggs

[aminal]

Apparently the built in flight simulator easter egg in OO is slow too.

More info on the source....

[zippity8]

Not to argue about whether or not OOo is more bloated than Office, but George Ou has always seemed to be ranting pro-MS and putting forth statements like this just to get the reaction.

Here's his webpage

And his other ZDNet entries

Also, you might want to check out the comments already posted to his review of OOo beta2

Lets see...

[Shads]

... open office being slow:

Java, now I'm no language bigot, but Java is slower than C (but more portable without changes in code).
It's a replacement for the most bloated piece of windows software and has most of the same features.

I use OO presently, it's not a speed demon thats for sure. However, A) It's free, B) Keeps me from having to run a windows emulator for word docs and scuh. So it's a win win. The equation would be ... 500$ vs lackluster speed, good compatability, and 90% features of office.

Re:Lets see...

[jdclucidly]

OO.o is written in C, not Java.

Re:Lets see...

[prof_tc]

Just one little thought.... Even with the somewhat questionable numbers of the 'tests', there wasn't anything over about 2.5 seconds. I don't know about you, but I just don't think that thats a long enough time to really effect productivity.

I can get numbers like that too, if I turn off quickstart. MS office also has a quickstart. You just don't notice it. I expect that makes a bit of a difference.

Re:Lets see...

[Anonymous Coward]

Browse to your OpenOffice install. Go to programs/classes.

What's there? JAVA LIBRARIES!

So, if it's written in C, why does it contain Java class libraries?

Possibly because, PART OF IT IS IN JAVA? Maybe?

Re:Lets see...

[fimbulvetr]

Acutally, it's written in C++. Also, it runs java on startup unless you disable it in options. Disabling helps quite a bit for startup time.

Re:Lets see...

[jdclucidly]

No. Only the JDBC driver from Java is used to provide database connectivity. Also, part of the forms wizard in 2.0 is written in Java. Everything else is C/C++.

Re:Lets see...

[sapped]

I just disabled it and it made a huge difference to the startup time for Calc. What functionality is killed in Calc by doing this?

Re:Lets see...

[fimbulvetr]

AFAIK, it only affects your ability to work with JDBC, so it's not like it should be enabled by default in the first place. The way I see it is:

If you know how to use JDBC, you're aware of how painfully slow anything java related is, and you're aware of how many hoops you'll have to jump through to get anything java working, so you should be prepared to go make some changes in open office so it starts jdbc at startup time.

If you are a newb to OO, or even experienced, yet have nothing to do with JDBC, why the heck should you have to waste 30 seconds of your life everytime you open up OO? It's such a dramatic difference, it's actually *bad* for OO to take that long to startup. You'd think they'd notice.

Re:Lets see...

[dslauson]

According to Wikipedia, here are the components written in Java in OO version 2.0:

• Parts of the Base application
• The media player
• Mail merge to e-mail (requires Java Mail)
• All document wizards in Writer
• Accessibility tools
• Report Autopilot
• JDBC driver support
• XSLT filters
• BeanShell, the NetBeans scripting language, and the Java UNO bridge
• Export filters to the Aportis.doc (.pdb) format for the Palm OS or Pocket Word (.psw) format for the Pocket PC

So true

[Arthur B.]

Honestly, I want to love Openoffice and to advocate it... I have worked in finance on excel, dealing with huge huge spreadsheets and many graphs... Have you tried to plot a 10 000 points graph in OOo Calc vs excel... in excel it is done in less than a second... In OOo the application will freeze for half and hour before slowly starting to display the graph. Cherry on the cake it will conviniently try to write "ROW" under each point in a huge ugly font. After that, changing the data means of course waiting half an hour again because the chart is updating. OOo calc simply doesn't do the job, how hard I wish it would.

Re:So true

[headCase]

I have also had this exact experience. OOo calc becomes completely unusable when trying to graph large data sets. I use gnumeric when faced with these situations and it is MUCH faster in the plotting.

Re:So true

[Arthur B.]

Actually I thought it was a bug at first... appears some bug reports have been filed for this... but in fact it's just the implementation that's like that... maybe to few people care about 1000+ datapoints graphs for OOo but this is still very weird. I can't figure what's taking so long.

Re:So true

[Yobgod Ababua]

My curiosity was piqued by this, so I went to try it.
Indeed, Excel can plot and manipulate 10,000 point graphs with relative ease, while OOo had to do some chugging (about 2 minutes for me rather than the OPs 30).

I wonder a little at why anyone would ever plot a 10,000 point graph in either program... all the applications I can think of are better served by graphing or scientific programs rather than a spreadsheet.

Useful tip for the OP: If you've specified your labels correctly (in an initial column for example) OOo will use those to label the X axis instead of the default "Row 1, Row 2, ...". By simply specifying a blank column you can get a nice blank X axis if that suits your data better. You can also change the font to something more appealing, if the default appears "ugly" to you.

Re:So true

[Arthur B.]

This has always been half and hour at least under every distribution I'v tried, debian, gentoo, kubuntu, fedora on various recent machines... I don't know how you managed to get 2 minutes

Re:So true

[J.R. Random]

I wonder a little at why anyone would ever plot a 10,000 point graph in either program... all the applications I can think of are better served by graphing or scientific programs rather than a spreadsheet.

Perhaps because he doesn't want to have to learn yet another application. After all, a spreadsheet can calculate stuff and graph the results, so he has decided to use his spreadsheet to calculate stuff and graph the results. Saying that he should have to move his data to another application to graph it is a rather lame excuse for not making OOo fast enough.

Re:So true

[Anonymous Coward]

I can guess why this happens.

You are plotting a 10,000 point chart in an area that is usually 500 pixels wide. Excel realises this and just averages every 20 points, rather than trying to display all 10,000. OO.o displays all 10,000. Theres a 20x speed increase right there.

Re:So true

[Arthur B.]

This was my guess too, but I think it explains only part of the truth... for example when displaying xy scatters where each point are clearly visible, excel still does it instantly while OOo takes an eternity.. I think it has to do whith the way it handles graphic display, cause other calculations are not that slower...

Re:So true

[ItMustBeEsoteric]

Wait...for professional functions you have to pay for software, but, for the AVERAGE, NORMAL USER, free is fine? Who'd have thunk!

Re:So true

[WWWWolf]

The horrible truth is that graphing, in general, tends to suck in all office packages. Back in the DOS era, I was a schoolkid and the only thing we had that could make graphs was Works, and it was really pain after a while (and this was just school project stuff, nothing as glorious as 10000 item plots); then, for me, came a long lull of not needing to do any graphing, and recently, now that I've had to fight Excel and OO.o, they've made the whole thing "easier" and almost impossible to use effectively for anything

Nowadays I keep going back to GNUPLOT if I want to graph something. Pain to work with (no "click and it does it"), but at least it gives understandable output without too much messing. Export stuff to CSV from OpenOffice.org Calc, let GNUPLOT shred it for a while, and it spits out EPS. Works perfectly each time, plus the graphs it makes have "dull scientific" look rather than "idiotic business" look, and I prefer dull scientific look any day =)

No kidding...

[Dr. Zowie]

The Mac port of OpenOffice (NeoOfficeJ) is so bloated that by default it starts up in the background when you log in! That's a crappy solution because it sits there hogging swap space until you want it.

I can start Mac Pages, Inkscape, Keynote and even the Gimp before NeoOfficeJ is finished loading. Now that's slow.

Re:No kidding...

[Master Of Ninja]

Sorry, what version of NeoOfficeJ are you using? I have the 1.1 version here, and there is no default start up in the background. Starting up NeoOffice J for the first time in a session might take 10-20 seconds, but each additional time to start up after that is quick (prob due to the java runtime needed).

Besides NeoOffice isn't a mac port per se. Its just re-implementing the X11 version of OpenOffice.org for the mac as a java application.

If you can start up all those programs before NeoOffice you either have something wrong with your setup or are a troll. Which one are you?

Re:No kidding...

[Dr. Zowie]

No troll -- maybe there's something wrong with my startup. I've got 1.1 going. I used the default install from the download (can't remember if it was an install wizard or a drag-and-drop) but somehow now it always auto-starts in my dock when I log in. The process is nefarious because it's not one of my usual startup items.

If I quit NeoOfficeJ and restart it, it takes about 20-30 seconds to start up on my PowerBook G4. If I work fast, I can click all four of the other applications in my dock and all four finish bouncing before my first NeoOfficeJ window shows up.

I'm running Tiger on the 1.67 GHz aluminum mac.

By contrast, Lyx takes 1-2 seconds to boot up :-)

Free RAM with open office

[clare-ents]

Well, according to the Misco catalogue I received this morning MS Office standard costs £300.
At my local computer shop, RAM costs £75/GB, so I could have 4GB of RAM for my machine.

On a price performance comparison MS Office uses 7MB and OO.org uses -3960MB.

Re:Free RAM with open office

[Anonymous Coward]

it's exactly this kind of idiocy that leads to slow memory hogs and the eternal hardware upgrade cycle.

if you can't write tight efficient code with proper use of resources, and can't design a system properly, then do something else.

Re:Free RAM with open office

[jabberwocky_rt]

While I understand your point... throwing more memory at the problem isn't going to really solve anything.

MS Office does a few things better, thats fine. OO, or someone else, needs to pickup the slack and improve their product, in both resource usage and features, before it will ever be taken seriously. Its an uphill battle, but certainly not impossible.

Re:Free RAM with open office

[Anonymous Coward]

Why only 4 Gb ?
Are you not feeling somehow "amputated" by having a 4 Gb limit ?

---
L'UNIX, 'cause I know what I want...

OpenOffice.org is not written in Java

[BenjyD]

Just to attempt to forestall all the Java posts - Openoffice.org is written almost entirely in C++, not Java.

Re:OpenOffice.org is not written in Java

[dozer]

But OpenOffice starts a lot faster, and runs faster, when you turn Java off. Hmmmm...

Re: So what's the FAX say?

[robw810]

The fix is broken too...
RW

XML Parsing

[Ca_computer_geek]

XMl parsing is memory intensive in general. So it does not surprise me that open office would be a little slower in parsing through the xml. Microsoft has been able to keep the APIs for their XML parsing closed to the general public. So if those API's ever see the light of day then Open office would greatly benefit from the code in those APIs.

Re:XML Parsing

[D3m3rz3l]

Ok, but then why can't the OOo developers create their own fast XML parsing code?

Re:XML Parsing

[Ca_computer_geek]

Well you would have to look at why they are using Java as the language of choice for parsing instead of something else. There are several question about this that I wish I could answer but even I don't know enough about this to give you the best answer on the subject.

Re:XML Parsing

[D3m3rz3l]

Well, they are probably using Java because

1) it's generally easier to debug Java than to debug C/C++ and
2) A lot of F/OSS advocates are in love with Java because of the magic phrase, "Platform independence".

Doesn't matter about the language's shortcomings. Personally, I think Java has it's place, but that place is not writing large, client rich applications such as an office suite. These days, I see a lot of people jumping to statements like "Java is *not* interpreted" and "Java is as fast, if not faster, than C++". I presume the reasoning behind such statements is the JIT compiler. However, a JIT compiler is still no match for a traditional optimizing compiler. Furthermore, even with a JIT compiler, the sheer number of run-time checks going on in Java would make it slow(er). Agreed, the run-time checks make for a more secure program, but I think that a sophisticated, experienced C/C++ programmer can also write a secure, modular, robust program. The Linux kernel is probably the best proof of that.

However, I've noticed a lot of people on this thread are stating that OpenOffice.org is not written in Java; That it's mainly C++. So I suppose the XML parsing is done in Java? C++ seriously needs better text processing libraries. Boost needs to become a part of the standard library. Then lazy people will not turn to Java/C#/Python when writing something like an XML parser.

memory is cheap.

[joey_knisch]

Again (already pointed it out on the blog) I want to point this out.

2G of cas 2 ddr3200 costs less than office 2003.

Bloats vs Bugs

[JPyObjC Dude]

MS has been very good at optimizing their software to run fast and have a relatively low memory hit *but* always at the sacrifice of stability and maintainability.

They have also had many years to optimise their code bases. OOo is stil very new and will get better over time. I doubt that MSo will improve.

Now don't get me started on the 1gb plus installation of MSO on my new laptop when I got it from my IT group :[

JsD

XML

[Anonymous Coward]

Most of that massive speed difference is due to XML being very processor intensive

So, why again do we have to use XML for everything? Couldn't you represent a spreadsheet just as well and transparently with something like s-expressions? And then you'd have a tight little parser in half a page of code, instead of a monstrous xml parser that handles entities, namespaces, etc.

What's even worse is using XML for remote procedure calls.

I know there's benefits to putting everything in the same basic syntax, but it's not like it's that hard to translate s-expressions into xml, anyway. With a little extra tool support you can interoperate fine, but keep it simple when simple is all you need. But use XML everywhere, following W3C's "all or nothing" rule, and your app has all kinds of extra complexity whether it uses it or not.

Abiword, Gnumeric, KOffice

[chill]

I've tried to use Open Office on my machine at home (dual-P3 800 MHz, 1 Gb RAM) and have always gone back to KOffice. OO has always felt "bloated" to me. It takes much too long to start up, and everything seems to slow down a little on my machine.

On the opposite end of the spectrum, Abiword and Gnumeric load very fast and seem to fly during use. KOffice is a touch slower than Abiword/Gnumeric but still light years ahead of Open Office. It also has a very snappy feel to it. Abiword works on Windows, Mac and Linux. Yes, I know, this doesn't address databases or presentation software.

IMHO, there should be no question mark, but more of an exclamation point.

  -Charles

Re:Abiword, Gnumeric, KOffice

[Anonymous Coward]

AbiWord and OpenOffice are bloated pigs compared to MS Office. See http://www.geocities.com/typopl/bug5291.html for a speed and memory usage comparison between MS Office 2003/AbiWord 2.2.X/2.4.X, and OpenOffice 1.X/ 2.X.

Re:Abiword, Gnumeric, KOffice

[Aim Here]

You neglect to mention that the comparison was a test devised by Microsoft, and involved importing a Microsoft-format file written by Microsoft about another Microsoft spec into a number of word processors. And appears on the homepage of a Microsoft employee. I think this may have some bearing on the result.

Oh, and the bug the test complains about doesn't appear on my copy of Abiword either, but I'm using a later version of Abiword on Linux. Perhaps it's just Windows at fault :)

Re:Abiword, Gnumeric, KOffice

[Anonymous Coward]

It might be off-topic, but you just raised a really interesting point: the whole concept of the all-in-one "office" package is a marketing creation/tool. It's about selling software to people who may only need one or two programs in the package, and creating the perception of value-- getting quantity at a discount while overlooking the fact that, say, your auditing department doesn't have any use for presentation software.

It's interesting that the commercial marketing concept of the "office productivity suite" persists in free software. Of course, the suite is now a convention, so it doesn't have to make sense anymore. There's a need to sell the idea of free software to potential users, and keeping with conventions can be a means to that end. But how much better served would users be by only installing Writer and Calc, say, if that's all they need -- and how much less bloat?

Re:Abiword, Gnumeric, KOffice

[Anonymous Coward]

Gee, you don't think the GP was maybe trying to astroturf? You were so close to seeing through his little scheme, but failed it at the last moment. Nice try though :D

To the GP: fuck you, asshole. Go back and squeal at your corporate masters. I hope you die a painful death.

Re:Abiword, Gnumeric, KOffice

[Buck2]

I haven't used Word/Excel in a long time, but I remember it being pretty nifty to embed a chart in a Word document and have it automatically updated when I changed the data points in Excel.

When you write up documents in multiple programs you need to find some sort of shared format (like PDF) within which to transfer plots and figures around. It's not always perfect, though. Converting an eps file to a pdf file is not always clean and there a seeming million other things which can go wrong at any given point.

The office "suite" should at least provide that sort of functionality.

Re:Abiword, Gnumeric, KOffice

[Geheimagent]

On the opposite end of the spectrum, Abiword and Gnumeric load very fast and seem to fly during use. KOffice is a touch slower than Abiword/Gnumeric but still light years ahead of Open Office.

Did you try to load the example spreadsheet from the article with gnumeric? It uses more memory than openoffice.org and it's slower. Saving the data and reopen it used more than 1.5GByte of memory before I killed the process.

Re:Abiword, Gnumeric, KOffice

[Anonymous Coward]

Geez, you didn't read the web page closely and you've got your anti-MS glasses on.
We were talking about OOWriter.
The test wasn't devised by Microsoft.
The page was created due to an AbiWord 2.0.x performance bug. I'm not a Microsoft employee. I'm willing to put money on that - you wanna take the bet?
I fixed a perf bug in AbiWord. Would a Microsoft employee do that?
If you read the later updates, there's a perf problem with the Windows version of AbiWord 2.4.0.
But that doesn't explain why AbiWord is still at least 3x slower than MSWord or OOWriter for this given test.
AbiWord also uses way more memory than either MSWord or OOWriter.

Re:Abiword, Gnumeric, KOffice

[Aim Here]

I don't know whether YOU are a Microsoft employee, Mr Anonymous Coward, and I never said you were. You did link to the homepage of someone claiming to be a Microsoft employee called Johnny Lee. See http://www.geocities.com/typopl if you don't believe me. Maybe he isn't a Microsoft employee,but anyone who'd fraudulently claim to be one is even less trustworthy on this subject, methinks.

"But that doesn't explain why AbiWord is still at least 3x slower than MSWord or OOWriter for this given test."

My point is that the given test could not be more Microsoft-centric if it tried. I don't dispute the results of the test itself.

"I fixed a perf bug in AbiWord. Would a Microsoft employee do that?"

So what? Raymond Chen contributes to the Linux kernel, and works for Microsoft. Go figure.

Re:Abiword, Gnumeric, KOffice

[Anonymous Coward]

I am Johnny Lee. I do not work for Microsoft Corp now.

YOU claimed that "the comparison was a test devised by Microsoft".

If you actually read the background of the web page, it all started from an AbiWord perf bug, #5291. Read the bugzilla entry for that bug, there's a link on the web page.

Your problem is that you don't like the data, so you try and discredit the source for the data.

Other people can corroborate the data. If you look at the timings, OOWriter 1.x was actually close to MS Word's time. Version 2.x is slower though. I and Robert Wilhelm reduced AbiWord 2.0's time from 92secs to 14secs for this test.

You said that because of these Microsoft-related points - "...this may have some bearing on the result."

But then you reply, "I don't dispute the results of the test itself."

Raymond Chen used to contribute to Linux, way back. I sincerely doubt he does that anymore.

Re:Abiword, Gnumeric, KOffice

[Aim Here]

"I do not work for Microsoft Corp now"

Right. that makes ALL the difference. I bet you've still got teh M$ cooties though.

"Your problem is that you don't like the data, so you try and discredit the source for the data."

Oh, I don't much care about the data one way or another. Maybe abiword IS bloated and slow, I care little since I don't use word processors much.
There does seem to be a lot of astroturfing about these days, so when I see something that looks like it, I investigate, because that sort of thing interests me. After all, I looked at your homepage, and see 'Microsoft Corporation' right there under your bleeding name n'all, and that's all hardly apparent from the link you pointed us at - what's a guy to think?

"You said that because of these Microsoft-related points - "...this may have some bearing on the result.""

Bah, picky picky picky. This is fucking Slashdot so you can excuse me if I'm not up to professional standards of lexical rigour, and maybe I phrased it without the level of precision required to pass an academic fucking peer review. And maybe this test was useful for solving Abiword performance bug 51421.44.X Plural Sector 12 alpha subsection 4, I know not. But that doesn't mean it's appropriate to use this small, narrow, and very Microsoft-centric test to make a general statement on the relative merits of Abiword versus OO.o verus MS Office, as you tried to do. What's next in this idiom? Linux sucks because it can't run Halo 2 very well?

"Other people can corroborate the data."
Perhaps the test is reproducible, I know not. My attempt to reproduce it failed, but I used a different platform, so I didn't dispute the results of the test. I'm certainly not calling you a liar and it's not me that called you an astroturfer. But from some angles, you certainly look a bit like one, you have to admit.

"Raymond Chen used to contribute to Linux, way back. I sincerely doubt he does that anymore."

His email address in the Linux CREDITS file ends in microsoft.com, so I was guessing that maybe he's added to the kernel since he joined them.

Re:Abiword, Gnumeric, KOffice

[chill]

Did you try to load the example spreadsheet from the article with gnumeric? It uses more memory than openoffice.org and it's slower. Saving the data and reopen it used more than 1.5GByte of memory before I killed the process.

Okay, I downloaded and tried that file on Excel 2003 (Win2K), OpenOffice Calc 2.0 (Win2K) and will try it later on KSpread 1.4.2 (Linux). Here is what I found:

The XML file is 189 Mb when unzipped. Wow, what a pig. Saving in native Excel format turned it into a 49 Mb file. Loading that into Calc and saving it out as .ods resulted in a 3.8 Mb file. Very compact.

Loaded into Excel, the Task Manager showed just under 50 Mb of memory usage. Calc was 210 Mb. I have no idea how much Excel is sharing with Outlook or other Windows apps. I need to do a clean reboot because the way Excel behaves on this machine is like it is sharing resources and the Task Manager memory numbers aren't accurate. (Note: Sharing is good, but I want an apples-to-apples comparison.)

The load time of the document into Calc was atrocious. Somewhere around 120 seconds. Excel was about 30-35 seconds for the first load. Saving in the native format of each was about the same as the load times. OpenDocument is an ordinary Java archive (JAR) containing standard XML files. JAR files are simply a set of files compressed together using the zip file format. This is why the file is so small, and I'm sure the compression/decompression contributes to load times.

Both Calc and Excel pegged the CPU at 99% for the entire document load time, seriously hammering the machine and making other running processes slow to a crawl.

Excel loaded the document much faster, and memory usage was seemingly better, but I think there is a bunch of "hidden" resources that are shared. I had Outlook open at the same time and had used Word earlier in the day. To be fair, I should do a clean reboot and try both again. Also try Calc after using Writer and see how they share resources.

However...

This spreadsheet has no bearing in the real world and this entire exercise it nothing more than jerking off. 16,282 lines of nothing but a few columns of raw data? No formulas, no references, no linked data, nothing. Then that data was copied over to 15 other worksheets. Same data.

This file was meant to do one thing -- make a big file. Period. Looking at 50 or so spreadsheets that are used on a regular basis where I work, it doesn't even come close to resembling any of them.

You need the right tool for the right job. If I saw a spreadsheet like this in production I'd tell whomever that they need to move it to a database, where it belonged.

This type of file (and blog entry) are why many people don't trust benchmarks. They have no reflection what-so-ever on real life usage.

I could just as well point to the 3.8 Mb file size of the OO.org file and say -- "Sure you wait a little longer loading, but you can't e-mail a 49 Mb (or 188 Mb) file to the client. You can a 3.8 Mb one. Even if you could, how long is a 188 Mb file going to take? It more than makes up for the difference in load speeds."

Do you know of any other comparisons that are more realistic? Every test I've done with stuff I use every day has resulted in "more than good enough" results for OO. AbiWord and KOffice have some issues that need to be resolved before they can be used more in corporate environments.

  -Charles

The true test of Open Source

[LexNaturalis]

The responses on ./ and the response from the F/OSS people will demonstrate whether Open Source is superior to Microsoft (or any closed-source company). If people just justify the results and claim that OO is still better just because it's Open Source, then in reality Open Source will lose. I think this is a time for the community to notice the problem, admit the problem, and then try to fix it. If the problem can be solved to the point where load times/memory usage is on par with Microsoft, then the Open Source community will prove that it is competent and able to produce a superior (or even equal) product that has the other advantages (freedom, lack of restrictive licenses, etc) that Open Source brings to the table.

Or... people can just whine and show the world that they're a bunch of babies who accuse people of being shills and just ignore the problem.

I, for one, hope the former occurs. I'll admit I'm not a good enough programmer (yet) to do anything about the problem now, but I hope the Open Source programmers who are capable will tackle this problem and fix it w/o making petty excuses.

Re:The true test of Open Source

[bobintetley]

If people just justify the results and claim that OO is still better just because it's Open Source, then in reality Open Source will lose.

It IS better because it's Open Souce. It's better in my eyes because I value my freedoms and I'd rather have a slower application with open formats and open code than a closed source application trying to lock up my data and tie me into one platform.

As you said, OO's problems can be overcome - Microsoft's can't.

Re:The true test of Open Source

[booch]

A few years back, there were some articles showing that Linux was significantly slower than Windows on 4-CPU systems. At first, there was some questioning of the results from the Open Source community. Once the results had been verified, the Linux kernel developers set about to remedy the situation. They were quite successful, and Linux has beaten Windows in every such test since.

Re:The true test of Open Source

[mc900ftjesus]

I hope you don't take the comments posted on /. as the F/OSS community's actual opinion. Plus, it's not like the "public" actually reads this site, and by public I mean people who have no idea what OSS is or what the point is.

Oh, and just diable java in OO, that fixes the load times. It's not OO's fault, blame Sun. You will lose the ability to run Java applets in you word processor, but who the crap asked for that feature?

Re:The true test of Open Source

[Cl1mh4224rd]

As you said, OO's problems can be overcome - Microsoft's can't.

That excuse is only valid in exceptionally small situations. To the average user open- and closed-source development is equally inaccessible. Your mother can do just as much about OpenOffice bloat as she can about MS Office bloat...

Re:The true test of Open Source

[alienw]

I agree. I would say the main problem is that OpenOffice is a huge, bloated source base. These types of huge, monolithic projects with a proprietary code base do not work at all for open-source development. It would probably take about a week to compile the damn thing on my system, and it would take me a lot more time than I have to get proficient with the architecture. Actually improving anything would be a challenging task, because the codebase is vast and is apparently of rather poor quality. To really improve anything like resource utilization, large parts of the codebase would likely have to be redesigned.

Closed source development is generally more efficient (thousands of well-paid full-time programmers vs. a few dozen volunteers), but open-source can generally deliver the same features with a lot less code, or at least distribute the workload among several smaller projects. The code generally tends to be better quality, simply because it's not rushed. With OpenOffice, none of these factors are true because it's an open-sourced proprietary product. We really get the worst of both worlds in this situation.

Re:The true test of Open Source

[BESTouff]

You got it completely backwards. "Open Source" is where it is exactely because it is open source, not because of some technical superiority. Freedom to use, copy, tinker with are what makes it successful. Moreover, open source can't lose because it has nothing to loose, whereas proprietary software "looses" often. OSS is succesful because of it's model, not because of it's technical merit du jour.

Re:The true test of Open Source

[GreyWolf3000]

s/loose/lose/g

Do you know what the 'g' does? That's right, it makes the substitution repeat, and I had to include it because you mispelt 'lose' more than once.

YOU FAIL IT!

Re:The true test of Open Source

[SharpFang]

Better? Defining "better" is one of the hardest tasks in science, no matter what domain. If you take speed as measure of quality, then yes. OOo is worse. Notepad is better. If you take amount of features, then Emacs is better. If you take functionality to price ratio as quality, then OOo is infinitely better than MS Office. (anything non-zero divided by zero...)
Many talk about freedom, okay, I don't give a shit about the product's freedom. I'm still going to keep using OOo because I can get it legally for free. I don't want to risk getting caught with illegal MSO, and a legal copy is more than half of my salary, and besides I never work with documents so big that the performance impact of OOo would mean any obstacle to me. The functionality is fine too. The worst thing I find is poor portablity of documents - sometimes layout of the presentation changes between computers - one uses slightly bigger fonts, a sentence gets pushed to the next page and layout of whole document changes and collapses. I need to be cautious about these and I find MSO superior to OOo in this respect. But not superior enough to justify the price difference.

Re:The true test of Open Source

[ObsessiveMathsFreak]

I, for one, hope the former occurs. I'll admit I'm not a good enough programmer (yet) to do anything about the problem now, but I hope the Open Source programmers who are capable will tackle this problem and fix it w/o making petty excuses.

You seem to have an advanced understanding of how open source works! You put up with the sub-par betas, keep on complaining to devs, and gradually things get better. Bravo!

Re:The true test of Open Source

[jacekm]

You have a full freedom to write your own office package and use it whenever you wish for free. You also have a freedom not to buy any M$ product if you don't like it regardless if open source exists or not. You have a full freedom to use whatever technically inferior product you wish. I grant you freedom to use open source free of charge "Trabant" plans to build your own automobile for free as you personal car instead of paying tons of money to the greedy BMW corporation for a ready to go product :-) If you really, really want everything to be open source, I suggest you to get familiar and eventually relocate if you find it interesting to one of those countries that do not appreciate greedy market economy (i.e. North Korea, Vietnam, Cuba, Bielarus). JAM

Re:The true test of Open Source

[Anonymous Coward]

FWIW, if you don't care about the database try to disable Java in OO. Then try measuring the startup times.

Re:The true test of Open Source

[bobintetley]

You sir, are a troll.

I outlined why OO is a better product as far as I'm concerned. My freedom and choices do not preclude you or anyone else from making your own.

Re:The true test of Open Source

[lkeagle]

He's not giving you an excuse. He's giving you a fact.

The real problem is not happening today. The issue is what happens in 10 years when your grandmother has all of her letters and documents stuck in a closed proprietary format, and MS has been found guilty of some huge accounting scandal and folds, no longer supporting Office 97? How is she going to be able to access all her files on her new computer her grandson just bought her because her old motherboard fried?

The answer is, she doesn't. It's lost. We have a word for this -- Encryption.

However, if the source for the document format was open, then all it takes is ONE enlightened programmer to produce a free tool to convert the documents to whatever new open format is currently the trend.

That's the argument, and it's NOT an excuse.

wtf

[Anonymous Coward]

OpenOffice Calc uses up 211 MBs of private unsharable memory

http://www.acronymfinder.com/af-query.asp?p=dict&S tring=exact&Acronym=MBS

So, Calc is using 211 Mass Broadcasting Systems? No no wait, that's 211 Millions of Bytes per Second.
That's fast!

ODS and Koffice

[digidave]

1. ODS documents are compressed (gzipped?), so large ones take longer to save and longer to open. I love them because I work with phone book listings and ODS files are small enough to email. Here's an example: 1/3 of the Hamilton, Ontario phone white pages is 51.8MB as CSV, 4.7 as XLS (Office 97) and 1.6MB as ODS. I also removed some unneeded data from the XLS version, so each listing is smaller. OpenOffice.org 2.0 opens the larger file sized XLS version five times quicker than it opens the ODS version, but ODS is still my format of choice due to the file size. Now, don't get me started on annoying 65,000 row limits... they really get in your way when you're dealing with phone books!

2. I use KOffice (KSpread) whenever possible because it's much quicker than OOo or Excel. I recommend it to anyone on Linux. Keep OOo around for some advanced stuff, but use KOffice for quick edits and browsing.

Re:ODS and Koffice

[Anonymous Coward]

ODS documents are compressed (gzipped?)

No, I think it's PKZip format, which is much better for extracting individual files from the archive than gzip.

Re:shoe on other foot this time? weird.

[Anonymous Coward]

Put the other way, extreme processing demands render progress in the hardware field almost moot.

ZDNet owned or funded by Microsoft?

[Anonymous Coward]

Is ZDNet owned or funded by Microsoft? Seems like every article I read from them is very biased towards anything non-Microsoft. Have you ever seen them do a pro Linux/Mac/Firefox/Open Source artice, ever? I am having a hard time finding one.

Use Gnumeric!

[Anonymous Coward]

I once had to deal with drawing a chart with ~40K data points. I had OOo installed, so I tried that ... bad move.
After strugling with it for 15 minuts, I went looking for an alternative. I found gnumeric, and what OOo could do in minuts, it'd do in seconds!

Gnumeric is full featured, and can save in compatible formats (for both MS Office and OOo). If you do any serious spreadsheet work under Linux, you owe it to your-self to try.

You can't just throw CPU at the problem.

[skids]

Give bloat an inch, next year it will be back for the mile. Needless IT upgrades cost companies money. It would be better if software folks stopped bloat. Then we could have faster CPUs *and* faster software.

That said, the only time I have found OO to be painfully slow, other than loading, is working with X/Y charts on datasets with more than 2k points or so -- it redraws every point like 5 times every time you mod the chart, and takes a good long time to do it. Too much conceptual abstraction going on there, for sure -- a pitfall of object oriented programming.

Re:You can't just throw CPU at the problem.

[isometrick]

Just a note: that may be a common error to make in OOP (and could very well be in OO.org), but it is definitely not a necessary one. Abstraction doesn't have to mean duplication of the same effort, and it can actually mean less of it.

Re:You can't just throw CPU at the problem.

[skids]

Absolutely. Just because the pit is there, doesn't mean you have to fall into it. :-)

must be on a windows box

[newSlashUser]

if i remember correctly, after compiling oo2, it ran very well and fast. the precompiled bins were def slower. my guess is that these tests were run on a windows machine. so just switch to nix and compile, its that easy.

Why not just evaluate the issue technically?

[Morgaine]

... an article possibly designed to sink Open Office

Maybe, maybe not, who knows. But what I find odd is that a simple, easily-measureable property like speed is treated as a religious issue and/or examined for conflicts of interest at all. Why not just measure it in a series of comparative tests as scientifically as possible?

And then, if Open Office is found to be lacking in speed, fine, no problem! The result simply becomes very valuable input to OO's design and development team, and in all probability will get dealt with very seriously and rapidly and to the benefit of its users.

There really shouldn't be an issue of contention here, if we're truly techies.

Re:Why not just evaluate the issue technically?

[arkanes]

The short answer is that speed is not a simple, easily measurable property. It's extremely open to subjective interpertation, which means that absolute measures can be misleading, and more importantly, it's extremely difficult to quantify and measure in an agnostic way. This is why artifical benchmarks are, at best, only casual predictors of real performance. However, they remain extremely popular, especially against those with an axe to grind, because it *looks* like a hard, concrete measurement.

All this is not to say that benchmarking isn't usefull, or that it cannot show real performance differences, but it's extremely naive to treat any benchmark as a concrete indicator of general case performance.

As an aside, in my experience OO.o often beats Office for speed, but in ways that won't generally show on a benchmark. For example, in one install (but not all - another reason benchmarks aren't generic) opening the Find dialog in Excel took about 5 seconds. Every time. Thats a very real percieved performance deficit. Office, especially Excel again, will often lock up as I resize or move columns around. This is a non-determenistic thing (related to Offices internal memory management, I suspect) and thus difficult to benchmark, but is real. It is extremely possible, even likely that someone in a different environment will see the opposite effects.

Re:Why not just evaluate the issue technically?

[theantipop]

So are you suggesting that we benchmark failure rates? I was always taught that benchmarks are meant to show average case performance in a well-defined environment. I don't think it's feasible to benchmark in a manner you seem to be suggesting. Should we benchmark the time it takes me to remember how to perform a column delete as well? It is important to remember that other factors can influence your performance, but speed as a benchmark should be as cut and dry as possible.

Re:Why not just evaluate the issue technically?

[arkanes]

Benchmarking obviously should be done in as cut and dry a method as possible, but it is *critical* to remember when publishing and evaluating benchmarks that they are synthetic and do not neccesarily generalize to actual performance. They are, at best, indicators.

This is especially important in an article like this one, where the author clearly has an agenda. For example, he characterizes (in his summary) OO.o as performing 98x worse than Office - clearly not supported either by the his benchmarks result (he's describing the worst case as the general case), or by good benchmarking practice in general - he's generalizing from one result on one system with one data set to a general case.

I don't want to imply that benchmarks are useless - they aren't. But *in this context*, they pretty much are. Certainly articles of this type, and in a publication like zdnet, are not the best candidates for objective performance testing. These are flame-generating hit-generators, even when the underlying basis (that OO.o is general case slower and consumes more memory than Office) are defensible.

With all apologies to Fark.com:

[Blakey Rat]

PWN3D!!!

Bloat vs. Cost

[no_pets]

With the money you'll save using OpenOffice over Office just buy a faster CPU and more RAM. It will speed up OO and still be money better spent.

The truth is in the middle

[LaughingCoder]

So everything this guy writes is suspect (i.e. wrong) because he ALWAYS takes Microsoft's side on things? Taking that thought a bit further, virtually everybody in this forum takes the anti-Microsoft side, so everything written here is suspect (i.e. wrong). I think that, as is almost always the case, the truth lies somewhere in the middle. Probably Microsoft is somewhat faster because it is C++ throughout, and it is a mature product that has had much optimization work done over a long period of time. It also probably uses less memory because it links with all those DLLs, and it is compiled C++. Conversely, OpenOffice is probably nowhere near as bad as the reviewer states; most likely he found an optimally bad dataset for OpenOffice to make his point.

Disable Java option...

[sarguin]

Go to the Options and uncheck the Java option (Use a java runtime environment). After this, OpenOffice.org start like a breeze...

Re:Disable Java option...

[KingPrad]

Thanks for the tip! I just did this and OO really is flying now. I can see it being usable now, whereas every time I've tried it in the past it was agonizingly slow. I'll definitely have to play with it more of then now.

Re:Disable Java option...

[sarguin]

I just don`t know what is the impact of disabling it... Probably some features need it.

Re:Disable Java option...

[ObsessiveMathsFreak]

I just don`t know what is the impact of disabling it... Probably some features need it.

The Sun Microsystems Logo?

Re:Disable Java option...

[tjw]

You need Java enabled to work with .odb files. Well at least any that use the built-in database which is written in Java:

hsqldb.org

I'm not sure if JDBC drivers are used for all external dbs, but probably.

Re:Disable Java option...

[hattig]

Yeah, you're correct. It works on NeoOffice/J as well.

13 dock bounces plus very very slow progress bar to 3 dock bounces and a decent progress bar.
(although I should reboot and do a proper test of startup times to be fair, I'm sure the 13 to 3 drop was merely having it in memory already).

If Java is only used for JDBC connectivity for the database component, then I think the default setting should be Java Off. In addition maybe Sun should do some work with optimising the start-up time of the JRE.

Re:Disable Java option...

[aminorex]

Yes, it is important to compare apples to apples. When I load OO Calc, it takes less than two seconds to start. I have Java turned off, because I don't need it. When I load Excel, Java isn't
an option, so any comparison in start-up times has to compare both systems without the feature
set supported by Java, otherwise I am comparing the start-up time of a system configured to include
Java extensions with one configured not to include Java extensions, which is inherently unfair.
On my system, OO Calc loads faster than Excel. That's just a fact.

Re:Disable Java option...

[mshmgi]

Wow - it's much zippier now. Thanks!

Re:Disable Java option...

[dimator]

I must be an idiot, I can't find that option in NeoOffice/J anywhere?

Re:Disable Java option...

[jiawen]

On my system, Java is disabled in Tools > Options > OpenOffice.org > Security.

Found this thanks to the excellent Open Office.org 1.0 Resource Kit, by Haugland and Jones.

Re:Disable Java option...

[aaronl]

Disabling Java does make OpenOffice start up and run much better. Unforunately, too many things in OOo are written in Java, and some of them are things that one wouldn't expect. The one that I just discovered was the MS XML import filter.

Re:Disable Java option...

[hattig]

it's at the bottom of one of the Options tabs, I forget which one. Ah, the Security one. (!?)

Actually, it only speeds up the progress bar part of loading. I don't think it is as much of a benefit as people have made it out to be.

Are you sure?

[Spy der Mann]

I downloaded OO.o the other day and opened a small ODT file. It didn't have any startup problems, and the interface (if ackward compared to the traditional MSWord) felt responsive.

So clearly it's something SPECIFIC to the Spreadsheet handling. And I recall OO.o being VERY SLOW opening spreadsheets in version 1.4 (or was it 1.1.4? whatever)

Positive M$ article On /. ???

[Anonymous Coward]

Wait.. an article on Slashdot that puts a Microsoft product in a better light than an Open Source product with no snide remarks tacked on by the editor???

Hmmmm....couple this with CmdrTaco suddenly learning to spell...

Penguins grab your pitchforks and torches.. I smell a Microsoft conspiracy here...

Appologists

[jone_stone]

There are a remarkable number of OO.o/FOSS appologists here. The answer to this surprising result seems clear to me:

Microsoft makes good software!

Okay, call me a troll, but I've tried a lot of free software over the years and I almost always find it lacking. Microsoft's stuff, on the other hand -- most particularly Office and Windows -- is remarkable when you consider how much they do and how efficiently.

One of the biggest areas in which FOSS is lacking is the boring optimization and debugging that's vital for world class software. The truth is that Microsoft is huge and has lots of money, so they can afford to spend time on that important finishing polish. There's an old saying in computer science: The first 90 percent of the work is easy, the second 90 percent wears you down, and the last 90 percent - the attention to detail - makes a good product.

Re:Appologists

[Anonymous Coward]

Microsoft makes good software!

Yes, from the people resonsible for:

Windows 1,2,3,95,98,ME,NT3,NT4,XPSP1,
MSDOS 2,3,4,5,
FAT,
Drive Letters,
LIM specification,
640k,
Thunking,
BOB,
Monkey Boy,
Clippy,
BSOD,
DLL HELL,
MS Works,
Notepad (with 32k text limit),
Outlook (any version),
Code Red, I Love You, Blaster,
VBRUN3, 4, and 5,
MFC3, 4, and 5,
IE3,4,5 and 6,
MDAC,
WinSock1,
MSWebCheck

Did I forget any?

Re:Appologists

[Khaed]

You just called Windows efficient. I'm not an apologist, but really, that's a little crazy. I find Windows lacking in every way -- it has awful permissions, it's bug ridden and a security nightmare. That said, Microsoft Office is pretty good. It's, hands down, the most well built software Microsoft puts out. But it does have flaws (the compatibility is a big one).

OO.o is pretty ugly compared to Word, but it loads .doc files better than Word does. OO.o saves .doc files at about half the file size. How is using twice the space "efficient"? This isn't anecdotal evidence: Copy this post into Word. Save the file as a .doc. The file size, up to the last sentence, is 24KB in Office 2003. For a paragraph and a half of text. Same exact file, opened and saved in OO.o? 9KB. Really efficient, there. If they can't even save a file efficiently... Of course, OO.o takes forever to start.

Important finishing polish... with all the bugs and compatibility issues and security holes? And you're accusing other people of being apologists? Pot. Kettle. Black.

Re:Appologists

[lkeagle]

Well, I guess we can't fault a company that puts in 270%

Re:Appologists

[marcosdumay]

Oh yes, MS has made some good software, it is called "MS Office". More precisely, it is called "Word" "Exel" and 'Power Point". Those used to be 3 excelent pieces of code (that MS is destroying now, but it is a slow process) and that should be acknowledged.

And, yes, FOSS developpers have wrote some ugly code. OOo used to an example (it is improving now, but it is a slow process), although some problems come from StarOffice code.

All the above is true, but I never heard that optimizations are boring! That is news for me, I tought programmers liked it so much that they needed to be educated to not optimize too early and too often.

sucky software

[Anonymous Coward]

Who said OSS can't suck? Just because OO's written by a group of geeks doesn't mean it's automagically better than software built by professionals.

Another Blog, another Bias...

[QuaintRealist]

As far as I can tell...

1) OOO IS slow - under Windows and Linux, enough so that competing "offices" like KOffice are kept alive despite reduced feature sets.

2) Office runs faster, but for that matter, so does IE - is it any suprise that MS can write software for its own OS which takes every possible advantage of its native environment to run with speed?

3) I use OOO whenever I can, because open standards means I know I'll be able to access my data in 10 years, unlike the struggle I've had with old Office/Wordperfect/XyWrite documents I've had to try to convert.

4) OOO is "bloated" in the same way my big multitool is bloated - you can't be small, fast, and everything to everyone on every OS

Re:Another Blog, another Bias...

[cp.tar]

These are the main reasons as to why I'd like to see a modular office package based on Enlightenment's graphical libraries.

I have a rather old computer, and although Linux itself is rather snappy - despite my terribly badly supported graphic card - OO.o is a pain in the posterior.

The new Enlightenment's (E17) libs look great - at least from my point of view, which is a designer's one, not a coder's one.
Unfortunately.

Anyway, I'm just losing my mind over a rather simple table that wasn't doable in the last 2.0 beta I tried - 'selected table cells are too complex to merge'. I'm still using 1.1, but I tried doing it on my father's 2.0 as well, and it still isn't done right.
When I recall how it was done in WP and how easy it has become even in MS Office, I must say I'm rather disappointed with OO.o, although it's a great program.

Re:Another Blog, another Bias...

[Anonymous Coward]

Answer to 2): My subjective feeling is that Firefox outperforms IE. Also Opera claims to be the fastest browser available, so it is apperently possible to beat MS.

Answer to 4): OOo is indeed huge, slow, and has much less features than Office. It does, however, run on several OS'es. (Office runs on Mac too, off course.).

What would have been nice was a smaller lightweight office package was LESS features, and a nice pluggable extension system like Firefox.

Re:Another Blog, another Bias...

[Urusai]

OOO is "bloated" in the same way my big multitool is bloated - you can't be small, fast, and everything to everyone on every OS.

MS Office runs on OSX, lest we forget, which is just a hop and a skip away from running on Linux. How does OOo on Mac compare to MS Office on Mac?

Re:Another Blog, another Bias...

[armentage]

2) Office runs faster, but for that matter, so does IE - is it any suprise that MS can write software for its own OS which takes every possible advantage of its native environment to run with speed?

That's a load of bull. There's nothing that MS can do to "magically" make their apps run faster on the same hardware that other developers can't, and there isn't much that Windows does that is significantly different from the way Linux/X do things.

Tricks that could make Excel perform better might be

• Leaving data on the disk until its needed on screen(using memory mapped files)
• Not copying data into temporary data structures for display
• Writing meta data along with the spread sheet to help it load faster
• Avoiding excessive repainting and better buffering of screen updates
• Writing their own widgets that work better in a spreadsheet than just using the stock ones that Qt/GTK/Motif gives you
• Trading speed for memory by using hashing and other complex data structures to navigate dense data structures rather than using sparse data structures that directly map to the screen.

What makes this notion especially naive is denying that Linux/X developers know their environment just as well as the coders at Microsoft. If anything, the longevity of Unix style OSes and X should make it far easier to write high-performance code, because everyone should be aware by now what the OS related pitfalls are. Regardless of the silly incompatibilities between the various flavors of UNIX, there really isn't much to differentiate them other than slight performance differences caused by how each OS pages memory or handles threads. Any one should be able to get similar performance on any UNIX, or even Windows, if they design their app's data structures well and are using compilers can do similar code optimizations.

The reason OpenOffice runs like a pig can ONLY be attributed to poor coding by its developers, period. Microsoft has had TWENTY years to hone Office into one of the most accomplished software packages ever. Thinking that they do this by using "evil tricks" is just ridiculous. It would be much easier to believe that the OO.o programmers are still new to what they are doing and have very poor quality control.

http://armentage.blogspot.com/

Re:Another Blog, another Bias...

[drsquare]

2) Office runs faster, but for that matter, so does IE - is it any suprise that MS can write software for its own OS which takes every possible advantage of its native environment to run with speed?

What advantages? You mean like not loading the whole suite every time you want to just use one part of it?

I don't see what's stopping the OO makers taking advantage of the environment in Linux for example.

Re:Another Blog, another Bias...

[jgoemat]

About #4: What "much less features" are you talking about?

Re:Another Blog, another Bias...

[CDarklock]

> is it any suprise that MS can write software for
> its own OS which takes every possible advantage
> of its native environment to run with speed?

If my software runs ten times better than someone else's, I tend to think it's because the other guy is a complete moron. The alternative, of course, is that I am an insanely great programmer... which I consider far less likely. There are not that many insanely great programmers, but there are an *awful* lot of complete morons out there writing code. So it is obviously easier for Sun to hire several morons than for Microsoft to hire several insanely great programmers.

Re:Another Blog, another Bias...

[MrResistor]

There's nothing that MS can do to "magically" make their apps run faster on the same hardware that other developers can't

Really? You've perused the Windows source code and verified that there are no undocumented APIs that give MS apps priority? No functions to the effect of if (author != "Microsoft") { run slow }?

Apps are dependent on the OS, and if the OS is Windows then there ARE, in fact, several things MS can do to "magically" make their apps run faster, and there's not a damned thing that other developers can do about it.

Re:Another Blog, another Bias...

[armentage]

See, this is the same sort of ridiculous Linux zealot garbage that I go out of my way to dispute. OpenOffice runs slower on LINUX that Office runs on WINDOWS. Microsoft can do nothing to slow down OpenOffice on Linux.

Re:Another Blog, another Bias...

[MrResistor]

I never said it didn't.

You claimed there was nothing MS could do to make non-MS apps perform worse than MS apps, and I refuted it. IIRC, that was one of the complaints Netscape made in US v. MS, though I couldn't tell you if it was upheld or not. I strongly suspect, though, that this was precisely the basis for the requirement that they open up their APIs.

99 seconds vs. 2 seconds

[Anonymous Coward]

Check out "passive noob"'s comments on OO start up times

This is crazy. I did a little experiment. JRE(Blackdown) on, CPU in powersave mode (800Mhz): 211 Seconds to start JRE(Blackdown) on, CPU in performance mode (1800Mhz): 99 Seconds to start JRE off, 800Mhz: 4 seconds. JRE off, 1800Mhz: 2 seconds.

I know Java fans don't like this, but Java is too slow for competitive apps. This is why it is restricted pretty much to serving as a learning tool and for interfacing with database packages.

Re:99 seconds vs. 2 seconds

[Glock27]

I know Java fans don't like this, but Java is too slow for competitive apps.

ROFL. Total BS.

This is why it is restricted pretty much to serving as a learning tool and for interfacing with database packages.

Yeah, and that's why the #1 Sourceforge project, Azureus, uses it? What about Eclipse, *the* most actively used and developed IDE on the planet? Eclipse startup is way faster than the times quoted in GP, and it's a) highly complex and b) written 100% in Java! Since no memory size is mentioned, I'd guess the GP needed more RAM in the machine.

Java has it's drawbacks, but they're not nearly as severe as you make out.

Re:99 seconds vs. 2 seconds

[aaronl]

How many copies of the JRE do you have in memory? How many copies of the C library, or C++ library? What's that? Oh, one JRE for each Java program, but only one copy of the C/C++ libraries. Hmm, that looks like a very bad thing for normal applications being written in Java.

Azureus is a good program, but not because it is Java. The developer was good at writing the program. It could easily be written in C/C++, and *it would perform better*. It would require less memory and it would have a faster startup time. Considering that the type of code doesn't lend well to runtime optimization, it also runs slower, in general, than an optimised native app. I run Azureus for lack of choice in good native Linux bittorrent clients. The program bugs me, though... the UI isn't very well done and isn't responsive. It also uses more memory than I would like, considering what it does.

Eclipse *IS NOT* the most actively used IDE... Microsoft Visual Studio is.

While I would say that Java isn't restricted to a learning tool and db interface, it is not used in a majority of desktop apps. When you consider how many desktop apps are *not* Java, it would be accurate to see that Java is hardly ever used there. Its most high profile use in a desktop app is also the focus of the current "slow and bloated" discussion, too.

Java does have its drawbacks, and I agree, they aren't as severe as a lot of people make them out to be. This doesn't make the problems inconsequential, though. I really isn't currently suited for desktop app development. There are problems that need to be addressed first.

Looking at my own system, if I instanced a JRE for each app I'm running, I would have used over 280MB RAM just for the JRE. (Est. 20MB per app.) That needs to be fixed in a big way before it is viable.

Re:shoe on other foot this time? weird.

[alienw]

I didn't realize how much I hated OpenOffice until I used Word for a while last night. OpenOffice takes about 20 seconds to start on my Linux machine. The latest version of word takes about 3, on a Windows computer with half the RAM and a slower CPU. I've not managed to crash Word in quite a while, while OpenOffice crashes reliably if you paste a figure from, say, Matlab and drag it the wrong way (I have about 20 of those Sun "thank you for your crash report" emails in my inbox right now). And god help you if you want to add captions to your figures, or use "styles", or insert an equation, or do just about anything a good word processor should let you do. As it is right now, I'd rather use Word under VMWare than a native version of OpenOffice. For now, my favorite by far is LaTeX -- even with its arcane syntax, it is a hell of a lot better than anything else out there.

That is not ad hominem

[A nonymous Coward]

Read this comment for a nice description of why that is not ad hominem.

Your slur on his 2 digit ID, however, is completely off topic. Google for "petard, hoist upon".

Re:That is not ad hominem

[jdclucidly]

I guess the subtle art of irony and sarcasm isn't as easily understood by the average reader these days... Did you really think that someone attacking someone for attacking the credibility of the source would turn around and do the same in the same breath? If so, I have a bridge I'd like to sell you... Have a little sense of humor, please.

Re:That is not ad hominem

[Anonymous Coward]

Wow you linked to an incorrect post to prove what? The initial poster is correct that the argument was ad hominem. Educate yourself a little instead of linking to
incorrect posts, sheep.

Re:That is not ad hominem

[dozer]

Too bad the post you linked to was wrong. Suggesting that a person might be wrong ("grain of salt" is just a nice way of saying this) because he or she has shown a bias in the past is most certainly ad hominem. Bias has no bearing whatsoever on a logical argument.

Old news

[archonon]

I've been saying this for years now and all i got bitching from fanboys. "No difference in my computer" Yeah right, i've installed OOo to 10-20 different configurations and where ever i test it, OOo is horribly slow to start. Slow startup is huge issue for many. Now you say that there is a quicklauch feature, but quess what? Maybe i don't wan't another memory hungry app to backround. (MS Office don't use any quickstart features either) Here's some numbers for you from my laptop. Coldstart: And memory usage (includes VM) OOo Writer: 31.6 seconds (~49 MB) Office Word: 4.8 seconds (~14 MB) "Warmstart": OOo Writer: 17.7 seconds Office Word: 1.8 seconds.

MODS - please mod parent down as troll

[Anonymous Coward]

parent post shows a link that claims to point to crossover office, but in fact points to the MS office main page. unless this is a typo, it's an obvious troll.

or is this part of MS's "Get the FAX" campaign?

OO copies rather than innovates

[vijayiyer]

I know I'm going to get modded down because I'm about to criticize one of the darlings of Slashdot, but I have some karma to burn: The problem with OO.org is that it attempts to be a clone of MS Office rather than a good office suite. I run MS Office using Crossover Office, and it is much faster (I don't have numbers, but it's on the order of 5-10x, even more for some very large documents) than OO.org. I'm not ideology driven - I need to get real work done. For me, that means document compatibility and speed. When I generate my own documents, I use LaTeX and Matlab to produce documents superior in quality to both OO.org and MS Office in an open document format. When reading large documents, OO.org is unstable, is not 100% compatible, and is too slow for day to day use. I tried opening one Word document killed OO.org after an hour, where MS Office opened it in a minute or so. Bottom line - why _should_ I be using OO.org? It suffers in compatibility on the reading side, and doesn't generate true publication quality documents on the output side. I understand that document compatibility issues are really Microsoft's fault, but, again, I have to be pragmatic. A couple of hundred dollars is negligible compared to the value of my time.

agreed, abi and gnumeric continue to be excellent

[Anonymous Coward]

both start up quickly and provide excellent functionality

Where's the data?

[Spy der Mann]

Dear Mr. George Ou:
We'd really appreciate if you provided us with the spreadsheets and documents in question, so we can run a benchmark and/or analyse possible memory hogs in OpenOffice.org.

Thank you very much.

Sincerely,
Your friendly neighbor Spy der Mann.

OO is slow.

[soupdevil]

I have OO 2.0, and Office 2003 on a Dell with 1gb RAM and a 2.66Ghz P4. I have the quickstarter loaded for OO, but not for MSOffice. Calc takes twice as long to load as Excel, Writer takes easily 3-5 times as long as Word on identical files. This is true when including starting the app itself, when the app is already open, the times are closer, but OO is still slower.

Re:OO is slow.

[magi]

I once begun to wonder what OOo actually does during the startup. I used strace to see how many files it opened or tried to open when I start and quit OOo. I counted some 1500 on one machine for OOo 1.1, 2500 on another.

Now with 2.0 (with strace -f), I seem to get about 2700 opens. This was just for starting and quitting OOo, and did not include opening an empty document.

About 1200 of the 2700 opens fail though, as if OOo was trying different locations to see if there are files for it. Still, there are 1500 successful opens. Actually, "grep -c 'No such file or'" gives about 2800 failed operations. I wonder why OOo likes banging it's head in the wall so much.

Don't know how long opening a file takes, but I guess it's not instant. On top of that comes processing the files.

You bet, in some cases

[dtfinch]

I have a spreadsheet that takes 15 minutes to open. About 20 rows and 3500 columns with a lot of simple formulas that took only seconds to recalculate in its entirety when the spreadsheet is already open, but for some reason takes 1000x longer to load when I close and reopen the spreadsheet, saying "calculating" in the status bar. I'm guessing there's at least one bug relating to this.

This is my only spreadsheet that gives me this trouble. I have many many larger, more complex spreadsheets that I work with in OpenOffice 2 without noticeable slowness.

No kidding?

[misleb]

Anyone who has run OO on Linux knows very well that it is bloated. Not only is it bloated, but it uses some homegrown toolkit for the GUI. I won't even use OO, personally. Normally I don't have a use for an office suite, but when I do I'd rather us MS Office, which isn't a problem because I now have a Mac sitting next to my Linux box. Of course, I have never paid for MS Office. Maybe if I had to pay for it I wouldn't use it.

I'm sorry, but OO is one of the worst examples of what open source is capable of.

-matthew

Re:No kidding?

[ccp]

Not only is it bloated, but it uses some homegrown toolkit for the GUI.

Right now I'm having OO Writer 2.0 and KWord 1.4 open side to side and they have the exact same icons and widgets.

I guess is a trick of the light.

Cheers,

Re:No kidding?

[misleb]

I haven't used 2.0 yet. Did they go full KDE? It would certainly be welcome. I know there is a project to add optional KDE support but I wasn't aware that it was anywhere near complete.

-matthew

Re:No kidding?

[ccp]

I haven't used 2.0 yet. Did they go full KDE?

At least in Mandriva 2006, it looks so.

OO 2.0 is much, much better than 1.1.5, do yourself a favor and try it.

Cheers,

Oops.. he DOES link the data!

[Spy der Mann]

http://www.lanarchitect.net/Examples/200264-l.zip

(And the obligatory coral link)
http://www.lanarchitect.net.nyud.net:8090/Examples /200264-l.zip

OK guys, time to benchmark!

"People [*don't] always look for what's cheaper ..

[Anonymous Coward]

Honesty is more important that cheapness. (Exceptions ... ???)

Not XML itself

[DV]

Saying it's because XML parsing is slow doesn't make sense. Any decent XML parser
will parse at a 30 MBytes/s rate on a recent processor, usually one waits for I/O
it rather how that XML data are handled that makes for a slow loading, not the
XML format itself. 2 minutes of processing would mean like a multi gigabyte XML
file, that's not the problem.

Daniel

mod parent up

[Anonymous Coward]

java is the problem, turn it off and openoffice loads tons faster

No, do not consider the source - fix it!

[arcade]

The author of the article had documents to back his claims. They may be specially tailored to make openoffice look bad - but I doubt it. In my experience, OO Calc is a pig compared to for example gnumeric - and that, quite frankly, sucks.

I only use OO to view documents various people send me. That's the *only* thing I use it for. When I write documents, they're either written in HTML (using quanta), or written in lyx (to make them pretty and printable! :)

Some of the software available as Open Source unfortunately sucks quite a lot. In my opinion, OpenOffice is one of those pieces of software. It's nice for compatability - but it sucks in usability.

Re:Bloat? Don't talk to me about bloat...

[dtfinch]

I think EMACS is closer to 50mb these days.

George 'FUD' Ou

[Foofoobar]

Every single article this guy puts out is heavily slanted, neglects to take into consideration loaded objects that Windows preloads and extremely anti-open source. I have yet to see him put out one article claiming anything that is open source is good.

If Microsoft isn't paying him, he must own stock.

Missing the point.

[elfguygmail.com]

This article is missing the point of OpenOffice. It doesn't take scientific results to find out OpenOffice is slower to open files than MS Office. Anyone using them will see the difference in a big way regardless what data they use. The fact that it's slow to open files and uses more memory is the primary issue with OpenOffice. The fact that MS Office can't retain compatibility accross versions is the main issue with MS Office. The point here is that OpenOffice is the first software in years that is on par with MS Office. They are both 'comparable'. People may argue on which one is actually better, but for the first time they are both pretty damn close in what they can or cannot do. Why is it important? Because on every single other aspect OpenOffice wins hands down. It's free, while MS Office costs hundreds. It's open source and totally customizable, while MS Office is closed. It's free of patent issues, while the state of Massachusetts found MS Office's proprietary format isn't. That's what counts. We finally have a real alternative, that is for ever free, and documents created by it will always be able to be read by any application implementing this open standard. So sure, shout out all day long that OpenOffice uses too much memory for your taste, but at the end of the day, it still wins.

Re:Missing the point.

[Valdrax]

Because on every single other aspect OpenOffice wins hands down. It's free, while MS Office costs hundreds. It's open source and totally customizable, while MS Office is closed. It's free of patent issues, while the state of Massachusetts found MS Office's proprietary format isn't. That's what counts.

No, it's not. Not in the marketplace. What counts is:

1) Can either application handle all your needs?
2) Which provides the best user experience?
3) Which has a lower TCO?

OO wins #3 hands down. It generally ties with MSO on #1, unless you really rely on spreadsheet macros (say, to handle timesheets). Memory uses & speed are definitely a penalty to OO in #2.

What matters in the market is how much emphasis each customer places on #1-#3. In general the markets don't care at all about patents and open source except as they affect issues #1-#3. You care a lot because it's a passionate issue for you, but it won't affect OO's acceptance in the market directly.

Re:Missing the point.

[PeelBoy]

Don't forget the fact that Microsoft Office costs several hundred dollars versus Open Office which is free.

That alone is worth a few seconds (mins top) of my time every day.

Besides I use Gnumeric not Calc. Let's see a comparison between Excel and Gnumeric.

Re:Missing the point.

[Anonymous Coward]

Why is it important? Because on every single other aspect OpenOffice wins hands down. It's free, while MS Office costs hundreds. It's open source and totally customizable, while MS Office is closed. It's free of patent issues, while the state of Massachusetts found MS Office's proprietary format isn't. That's what counts.

None of that is important to me. I just wanna write. OO.o is useless for that.

Re:Missing the point.

[Anonymous Coward]

In fact, the article is pointing out that they are NOT comparable.

Re:Missing the point.

[Anonymous Coward]

None of that is important to me. I just wanna write. OO.o is useless for that.

May I suggest a pen?

bloated compared to what?

[SaidinUnleashed]

OOo is VERY bloated.

At least when you compare it to MS Office 97. MS O97 was FAST.

Compare it to MS Office 2k or 2k3, and OOo is MUCH faster, at least in my experience.

Dollar sign causing error?

[Overly Critical Guy]

Hmm, I wonder if it's the dollar sign in the file name? Perhaps renaming the file will correct your problem.

Load Times

[Bazman]

His table compares the time it takes OOCalc to load an ODS and SXC document with the time it takes Excel to load an XLS (or XLS-XML) document. Now that's not the same thing.

He should compare the time it takes OOCalc to load an XLS document with the time it takes Excel to load it. Sure, maybe Excel wins here by a few seconds.

Now compare load time of the ODS document in OOCalc, which takes three minutes (it says there) with the time it takes to load it into Excel. I predict that will be at least three years, if ever...

OO Wins!

Baz

The Windows OpenOffice version is bloated

[Anonymous Coward]

The tests were done in Microsoft Windows, so the test is biased.
No Linux based test was done for Microsoft Office, so, since Windows is MS Office's optimized environment it will obviously do better there.

The same logic applies to OpenOffice. No optimization in the environment, worst results.
A Linux test would show OpenOffice running better than MS Office, since that is it's optimized environment.

The test is not sound, and it couldn't be. Not at this time, since no unoptimized version of either works in a neutral Operating system.

What a joke

[Anonymous Coward]

When MS does something right and better, it is because MS cutting corner and cheating. When OSS does something bad, it always has some justifiable excuse. This kind of attitude is not going to be good for OSS.

Re:"Essentially" the same data? - MOD PARENT UP

[haruchai]

That's a good point - money spent on better hardware is better than on a single piece of software - unless you don't have an alternative, of course.

External XML

[Doc Ruby]

XML is a data interchange format. We've finally arrived at the era where apps can interchange data in XML without (necessarily) being trapped in a proprietary data format. But that doesn't make XML suitable for internal data representation. Apps should use internal data formats that support their native performance, and serialize data objects to XML for interchange, including storage. Using XML internally when performance thereby suffers is the bad kind of lazy, bad design that saves development time at the manifold expense of user time.

The one key difference

[squoozer]

I don't want to be an apologist for OO but you can't deny the fact that MS has had about 10 years long to get MSO right than the OOo people have had to get OO right. Now that isn't to say that we should or will have to wait ten years till OO is as good as MSO is today but we should cut them a bit of slack if the software isn't a slick an lean as it could be. In a very short period of time the OOo team have gone from nothing to something that can rival MSO. Assuming the pace of development continues OO will, I feel, be as good as MSO in two years.

Re:The one key difference

[Anonymous Coward]

Also, don't forget that MS has invested billions of dollars and supposedly the brilliant minds of hundreds of programmers on Office.

Re:The one key difference

[dozer]

You can't deny the fact that MS has had about 10 years long[er] to get MSO right than the OOo people have had to get OO right.

Are you joking? StarDivision was founded in 1986, and some code found in OOo goes back almost that long. StarOffice was created in 1994. Depending on how you count, I would say that StarOffice and OpenOffice are within a year or two of each other in age.

Two years until OOo is as good as MSO? You're dreaming! I'll take that bet.

Personally, I use Gnumeric for all my spreadsheet tasks, and I eagerly await the day when Abiword doesn't randomly crash when a document contains footnotes.

Re:The one key difference

[HikingStick]

While I agree with you in principle, I do believe it fair to note that the level of difference in response times and resourse usage (the size of the gap) is significant. I have no doubt that the OO community can come to grips with this and streamline the app, but it does highlight a potential pitfall for the Open Source commnunity at large-- unless the project plan establishes specific performance levels up front, there is no specific target towards which to work. It needed not get this far before taking steps to address it.

What is truly "bloated"?

[nmaster64]

The only thing I see bloated here is Microsoft's wallet...

Microsoft Office: $10 per 1 second less load-time

OpenMediaOpenOffice.org - slow, bloated?

[Anonymous Coward]

Here's a link that contains recordings from an openoffice conference. One of them talks about the slowness of openoffice.

http://ooocon-ljudmila.kiberpipa.org/media/

OO isn't the best office

[Qbertino]

OO is only better than MS because you don't have to pay for a bloated mess. The best office package I've used is Lotus Smart Suite. I'd be glad to pay a three digit sum for a cross-plattform version (Linux/OS X/Win) of that office package.

Re:Bloat? Don't talk to me about bloat...

[The Famous Brett Wat]

I have compiled OpenOffice from scratch... took a while!

I realise you're trying to make a statement with all your telnet to port 80 instead of using a browser and such, but I think you should at least have used a compiler to build OpenOffice.

My own timings.

[Shayde]

So this is an interesting thread, and I thought I'd take a look at my own situation. There's something odd going on here, and I'm not sure what it is in OpenOffice that does this.

I started OO on my Debian Linux (Sarge) Thinkpad T40 (my original article on this) while doing my normal workload (Evolution, IRC, etc etc). As soon as the main Frame appeared, I clicked the [X] close button to exit it - assuming that OO's main startup thread would have completed, and shutdown processes are pretty fast. My first timing:

dbs@hunter:~$ time openoffice

real 0m30.440s
user 0m3.084s
sys 0m0.216s

This is not a particularly fast laptop (1.4gig), but not bottom of thel ine either. That's not too bad to start up a whole office suite.

But, I decided to try a few other things, so from the same command line, I did it again. It came up VERY fast:

dbs@hunter:~$ time openoffice

real 0m3.475s
user 0m3.104s
sys 0m0.128s

Nothing in 'top' is showing memory cached or in use or anything. I know that Microsoft uses a 'quickloader' to keep things like Excel and Word in memory between uses, so there's a memory hit in favor of faster startup, but I wonder if OO does some sort of environment checkup on start each time, and remembers setting between restarts (in a gconf config entry maybe?)

Now, 10 minutes later, startup is still only 3 1/2 seconds or so.

There's so much mystery and hand-waving in Microsoft products, that benchmarks like what this guy is talking about are nigh on useless, and saying that OO is 'bloated' is playing to the knee-jerk folks who constnatly rail "BLOAT IS BAD!". Bloat means useless functions inside a product that have no end-user need, and are added to the size of an application in order to appeal to a very very narrow audience (or just due to bad design). I don't think OO is in this category at all.

As far as the filesize loading problems, I have yet to try this out, but I shall, and will report on what I find.

Re:My own timings.

[dmaxwell]

The Linux kernel is very aggressive about caching and buffering memory. I suspect the subsequent relaunch performance would drop a bit if you loaded up and stopped a few other large apps or even better a few large files in Gimp. That would kick a lot of those cached memory pages out of the caches.

Office Suites?

[Anonymous Coward]

I use notepad.

OO memory usage

[bperkins]

OO is pretty usuable, as long as you have a very large amount of RAM. I upgraded my thinkpad from 256M to 1G and openoffice load times went way down ( probably ~5x under some circumstances).

I'm not at all familiar with the architecture of OO or what the developers priorities are, but it'd be nice if a bit more time was spent on performance. Firefox could also use work here also.

I'm sure that OO wants to concentrate on features and compatability. That's certainly a worthwhile goal, but perfect compatibility seems pretty much hopeless, and you can always think of more features to add.

Re:OO memory usage

[British]

OO is pretty usuable, as long as you have a very large amount of RAM. I upgraded my thinkpad from 256M to 1G and openoffice load times went way down ( probably ~5x under some circumstances).

So instead of paying X dollars for MS Office with a faster startup time, you are instead paying X dollars for open office to start up equivical to Office's.

Perhaps everything isn't cheaper in Linux land. Instead of paying X dollars for a closed-source product with a nice printed manual, you get an open-source version with no printed documentation, but instead buy an O'Reilley book with a cute animal on the cover. So you have to pay either way. And here comes a hundred technical workaround reasons why OO is slower from OSS advocates.....

Re:OO memory usage

[bperkins]

I didn't upgrade to run OO, which I only rarely use. 256 wasn't quite enough to run the desktop, firefox and say acroread to run smoothly. It also makes apt-rpm run about 5 times faster as well.

OO being much faster was just a side effect I noticed.

Besides they did't sell Office for Linux last time I checked. Money isn't the issue, I just prefer Linux. My comment was an observation, not a recommendation.

Compare Filesizes?

[gQuigs]

The Microsoft's XML size is 188 MB(197,521,133 bytes). The OpenOffice's SXW size is 3.59(3,770,916 bytes). Would you prefer to email a 188 MB file or a 4 MB?

Lotus 1-2-3

[dalewj]

I miss you, where are you?

file size is larger too

[hobo sapiens]

I like OOo and recommend it to any of my friends who will listen. I got v2 and still like it a lot, even though it is still slower than Office (opening, saving, etc). Jeez, wonder if it has anything to do with java? Of course, I have been flamed previously for suggesting that java is slow, so I will say that java is the best ever. It's neato and way rad. And super and stuff like that. Right.

Anyhow, one interesting thing I noticed is that when I converted some docs to the OpenDocument formats, the file sizes are nearly twice as large. Not that document size really matters anymore (my cheap PC has an 80G HD, and yes, I left the door open for someone to reply to this and get some funny mod points with that last stmt and closed it with this stmt). It does make me wonder, though, if part of OOo's sloth is that it is rendering / creating XML documents rather than the old file format? If that's the case, I am sure in time it'll get better. Plus, if M$ were to ever adopt OpenDocument (riiiight) I'll bet they'd have the same problems.

Actually.

[Agarax]

This is one area where Open Source has its weakness.

Cutting down and optimizing existing code is not nearly as glorious as adding new features.

Micro$oft, on the other hand, can afford to have a whole team of programmers who's only job is to optimize and slim down the code.

As much as I hate MS, they did get a lot of things right in Office (except for that damn paperclip).

Re:Actually.

[Philip K Dickhead]

Microsoft - for all of its size - has only about 8,000 FTEs on development teams.

Re:Actually.

[Anonymous Coward]

I manage a team of programmers and let me tell you, the most praise and glory gets heaped upon them when they can optimize a program that's been acting like a pig. Idunno, maybe that sort of work isn't going to get you a mention in the Wall Street Journal or even CNet, but optimization is SO much more important than people realize these days. Let me offer you a nice, well-rehearsed rant about the quality of programmers who have graduated since say, 1987... Oh, forget it. I'm sure somebody else has already posted one. I'll just say wise up, Sonny-boy. The next generation of CPU and memory WON'T catch up with your crappy, wasteful programming practices. Show some pride of craftsmanship for God's sake!

Re:Actually.

[Deathlizard]

I don't know if it can be said it's a weakness of open source as of yet. This openoffice is still relatively new code. It will probably get some patches that will speed up functionality over time.

Yes, Microsoft has the programming resources and capital, but they never seem to use it for streamlining except for once in a blue moon. I think Office 2003 is the only time I can think of that they actually released a product more for streamlining than for functionality, and what resulted from it is one of the best running offices in a long time.

There really isn't much of a difference functionality wise between office XP and office 2003 (outside of outlook), but there is a big difference when it comes to size, where office 2003 fits on one CD, Office XP takes about 3-4. (1 for most office apps, 1 for clipart, 1 for publisher and an optional 1 for a tutorial CD.) They even added infopath and onenote to office 2003 and it's still fits on that single CD. On top of that, it's more reliable, faster and easy to migrate people familiar with XP.

Added functionality is one of the things that gets Microsoft into trouble a lot. They love to add functionality instead of streamlining their product. The next office and windows just looks like their going to be a mess from the pictures I'm seeing. I really wish that Microsoft would just keep streamlining their products instead of reinventing the wheel every three years to try and get that Whiz! Bang! Wow! reaction they got when Win95 was released.

Re:Actually.

[Mugros]

Well, actually... Bill Gates was the one who said that they would rather implement new features than fix old bugs.
That is why every year or two a new office version comes out, but basic features are not improved.

I am working with OpenOffice and i am migrating my documents from MSO to OOO as needed. As OOO is V2.0 i would guess that it will be improved over time. Full build bumbers are always not very optimized.

Re:Actually.

[Anonymous Coward]

As a professional programmer and open source developer, I must tell you that you are wrong.

Optimizing is actually quite fun usually to do IMHO, but the reason why it isn't done so often is priorities and time. For example, which do you think is worse: Corruption of a text document when saving or using 5MB of more memory than is needed? Or crash when program is started or 20% speed loss when document is saved? Developers would love to do things they like, but they have to follow priorities and fix the worse problems first. Sometimes a lack of a feature is a worse problem. So currently they have bigger problems that memory usage, but of course they have to fix problems in that area at some point also.

Re:Actually.

[johnnyb]

"Cutting down and optimizing existing code is not nearly as glorious as adding new features."

Bzzzt. Wrong.

The problem with OpenOffice is that it started out as a commercial product, and has been trying to fix itself ever since. This was nice because it already solved the problem -- an Office-like suite of tools -- but it was fairly braindead in the way it went about it. But that's what we inherited, and so that's what is being worked on.

If you want to look at open-source projects for a development style comparison, look for the ones that were open-source from their conception, like AbiWord and Gnumeric. Both of these apps are extremely light on their feet, and very good. They both still have some feature catch-up to do, but I am confident that they will.

Re:Actually.

[ashyanbhog]

MS Office=Slimware : OOo=Bloatware is not true. Going beyond the numbers, I recently had to open a 150 MB PPT file full of graphics, and OfficeXP displayed the file within 10secs, but as I began scrolling down, the whole machine just slowed down as Office began rendering the slides one by one, and at a point, GUI stop responding After killing office, I opened the same file in OOo2 (which was still in Beta release back then) and to my horror it took close to 15 - 20 mins to open it....... but after the file opened, OOo was running as usual, and editing the file as as snappy as editing a simple 2 or 3 slide PPT. I am product designer just out of college, and when I was there we were preparing our assignments in M$ word and used to used to often notice that Word would begin screaming after a few graphic files and heavy formatting was inserted to word. OOo2 at home would have no such problem and as it matured with each test build, I switched to creating my assignments in home and soing a final edit at college before submission.

Re:Actually.

[Anonymous Coward]

Even though OOo works nicely it is really slow. Starting up Calc on a 1.6 GHz Pentium 4 takes approximately 10 to 15 seconds. Excel is a lot faster than that on a laptop with a similar processor. I think that this doesn't necessarily have to be a problem when doing stuff at home (because I prefer free software) but during work I want things to be fast. Saving a 150 KB OpenDocument Spreadsheet in OOo Calc takes also about 10 seconds ... Plus this doesn't happen in background but locks up the interface. Why?

Re:Actually.

[Anonymous Coward]

This isn't really an open source problem. More like fixing a problem that existed when the software WASNT open source.

Don't forget Open Office wasn't always free and wasn't always open source. It's based off StarOffice.

Open Office is MUCH better now than it use to be and it's only getting better.

Re:Actually.

[Anonymous Coward]

Micro$oft, on the other hand, can afford to have a whole team of programmers who's only job is to optimize and slim down the code.

At one time I worked at Microsoft, and I worked on part of Office. My experience there does not support your idea.

Perhaps, MS can afford to do this, but they do not.

Here's how development works at MS:

1. Program Managers decide on a list of new features for the next version. They consult with development leads as part of this process (that is, devs can suggest features, but most features come from PMs).

2. Developers code up the new features.

3. Testers find bugs.

4. Developers fix bugs.

Almost never is any work done on making the products slimmer or use less memory. Its all about the features.

When I was there, my experience was that MS developers were good at micro-optimization but not overall. If an MS dev was writing a new feature it would probably work and would probably be reasonably efficient. But they spent almost no effort on code reuse, or reworking existing code to slim it down. And how could they? They had all these features to code up, and were already working 50+ hours per week to try to make their schedules... and slimming the product was not a scheduled item.

I remember that there were two functions: one to lookup a string in a data structure, and another one that did the same thing but was case-insensitive as to the string. You might think there was a core function, with an argument saying whether to be case-sensitive or not, plus maybe a couple of wrappers. Nope, there were two complete whole functions, almost identical. Worse, the calling conventions were different: one was FLookupFooI(sz, pfoo) and the other was PfooLookup(sz). In English rather than Hungarian, one took a pointer as an argument and assigned the result through the pointer, while returning a bool true/false flag saying whether it succeeded; the other passed back the result and had a special result to indicate failure. So to change code from case-sensitive to case-insensitive you had to rewrite the call completely, rather than just adding in the "I" to the function name.

MS applications represent man-centuries of effort. There are lots of little hacks that are not really understood now or remembered, and MS is worried that if they start making big changes to slim things down they will break something that was already fixed long ago. As long as things are "fast enough" for current hardware, they won't change this.

Re:Actually.

[synthespian]

AbiWord has a plug-in architecture. That's really promising, IMHO.

Re:Actually.

[LaissezFaire]

I've worked for a couple of startup software companies. Speed is a feature, just like writing in the OOo format is a feature. Companies (or open source software projects) that stay in business provide the features that their users want. Right now I'd say that MS Office users have wanted the speed feature more than the OOo format feature. OOo users are still getting some basic MS Office features, and that has trumped the speed feature, at least for now.

Re:Actually.

[Anonymous Coward]

"This openoffice is still relatively new code."

Huh? So when OOo 3.0 comes out, comically bloated as ever, you'll say "oh, it's just new code". And then for 4.0...

OpenOffice.org and its predecessor StarOffice has been in development since the 1980s. Yes, much of the code has changed, but since it started to mature around SO 5.2, they could've done loads of code cleanups, and making the thing remotely elegant and efficient.

You can use your argument forever, and it's still utterly meaningless.

Is this really news??

[kilgortrout]

MS Office is more efficient at handling its own closed format files than Open Office 2.0. Who would of thunk it?? Look at the time differences. Apart from the specially crafted spread sheet, we are talking about tenths of seconds for MS office as opposed to one or two seconds for OO. Generously assuming the average worker opens maybe 100 documents a day, we're talking less than two minutes a day difference. People waste more time than that scratching their buts. I'm not saying this is not an issue which should be addressed by the developers of OO, but to sensationally report that OO is seven times slower without giving equal emphasis to the number range you are really talking about(tenth of second vs. 1 second)smacks of the usual marketing FUD from Redmond. Also, there is no indication that any care was exercised to insure that the documents used in the test are representative of the types of docs encoutered by most people. I strongly suspect they were chosen to cast OO in the worst light possible. I have opened many .doc documents with OO 2.0 and did not notice any significant time lag(less than half a second for typical 100KB .doc file) although its obviously a little slower than MS Office. So what you have here is some self appointed blogger with an apparrent axe to grind against open source posting his supposed test of OO and MS Office in a skewed and sensationalist manner and with results that are contrary to any real world experience or concerns. This is not news; it's marketing FUD.

MS Office's quickstart

[lordscotus]

OO writer starts in less than 10 sec's on my K7 1.4GHz with 1G, and that's with no quickstart. (Disabling java only cuts that marginally.) I can live with that.

Remember, that the entire windows OS is the "quickstarter" for MS Office. All the dlls and IE stuff are already loaded.

So...

[design.sound]

...Comprehensive ...Freedom ...Working

Pick any two.

GCalc?

[Anonymous Coward]

Nobody has mentioned gcalc yet? It's about a billion times faster that OO Calc and has many times more useful features.

Why they don't care

[gr8_phk]

"If it's legally free to use and does the same task, why wouldn't 90% of the users in the world who only use Windows *not* care?"

Because they don't care about "legal". Often when I tell someone about OpenOffice, they tell me it's neat but they already have MS Office at home - or at least word. If you tell them "but it's free", they often say they got the MS products free too - illegal of course. They figure why get some free knockoff when they can get "the real thing" free. The ones who paid for MS often got a student price or something, and they really have no incentive to switch until their existing version won't work any more.

The problem is that everyone has Word or Office already weather they paid for it or not. In that context, OOo has nothing to offer - the other benefits are too abstract for joe sixpack. It's a case where MS benefits from casual copies floating around.

The situation is the same for others: Mechanical Engineers tend to have a pinched copy of Autocad at home. Artists have a pinched Photoshop. Animators have a pinched copy of Maya. This hurts adoption of GIMP and Blender - sorry, there is no great GPLed CAD program (except for QCAD for 2D). I'm sure there are plenty more examples. If Longhorn can prevent people running illegal copies of all this software, we'll start to see people switch - assuming MS will allow them to run the legally free stuff.

Re:Why they don't care

[level_headed_midwest]

Here's why more people don't use OpenOffice: 1. THEY'VE NEVER HEARD OF IT. Most people don't know jack squat about computers or programs. They use what everyone else does or what they've seen elsewhere. That would be MS Office because that's what they have at school or work. They don't know that there are any other office suites even out there. 2. If they do know about OpenOffice, they don't like it because if they had ever used it before, the commands are in slightly different places on the menus than in the version of Office they use at work. This made it "too hard to use" because they have to re-learn a few locations of functions. (Interestingly enough, most people I know HATED Office 2003 when it first came out because the commands and menus were a little different than in Office 2000. They said it was impossible to use! Same thing for Windows 98 users than went to XP.) 3. They opened up the most heavily-formatted Office 2003 document they could find- lots of macros and such. It didn't open up quite right in OpenOffice, so they concluded that it was junk, never minding that Office 2000 or XP would have barfed on it worse. I used my Linux-running computer to display a read-only PPT 2003 presentation off of a USB stick after the presenter's computer crashed. He was SO pissed that one hyperlink didn't work right (linked to a non-existant file on the "E:/" drive, but the rest of the presentation was *perfect.*). So he used somebody else's computer with Office 2000. The text boxes were all over the place and his background was gone when they displayed it... 4. You can get MS Office for free from peer-to-peer or by sharing an original disc.

Re:Why they don't care

[swimmar132]

I would think that large corporations and government agencies generally care about the legality of their software. So, you might see most of the initial movement towards OO there.

Re:Why they don't care

[jimicus]

Large corporations and governments also have thousands of existing documents to convert. And when you're a large enough organisation, there's no such thing as "list price".

"Free" is all well and good, but "fails to open a significant percentage of documents and half of our staff point-blank refuse to use it without training" suddenly looks like a big problem. Very few businesses can successfully sack 50% of their staff ;)

Re:Why they don't care

[Kluge66]

Well, there's BRL-CAD (www.brlcad.org), although I imagine it's a very different beast than AutoCAD.

WTF? Somebody's full of it.

[slackmaster2000]

I was a little concerned that Calc might use up 211MB of memory, but it didn't sound right so I checked it myself.

The OpenOffice binaries on Windows use up just over 40MB when OO is "closed" (e.g. the quickstart program is still in memory).

When I launch calc, that number shoots up to an astronomical 42MB. Heh. Opening a few fairly simple spreadsheets in MS Excel format and the number goes up to 60MB.

Excel (2002) on its own uses 15MB when it's running empty, and opening the same few fairly simple spreadsheets jumps that number up to 21MB.

So yeah, Excel is smaller in memory. I also find it to be much more responsive. But Calc is not 600% bigger for crying out loud.

Anyhow, sorry for the lack of details but I don't feel like going into detail. I'm on XP SP2 with 1GB RAM. Office XP 2002 and OpenOffice 2.

Re:WTF? Somebody's full of it.

[slackmaster2000]

Oops, I should know better than to trust article summaries.

The summary lead me to believe that the author claimed that Calc had a 211MB footprint. This is not the case.

My bad.

Re:Bloat? Don't talk to me about bloat...

[aminorex]

Eighty Megabytes and Constantly Swapping.

Emacs grows to fit the available space. Or at least it did when it was still growing.
It seems all the competent hackers have lost interest, because everybody has been
trained not to understand lisp.

Poisoning the Well

[Ksisanth]

Circumstantial ad hominem (poisoning the well) is using the particular circumstances of the messenger to attack his message. It is indeed a fallacy, even when the claim about his circumstances is factual. Specifically, it is a fallacy of relevance, because those circumstances are irrelevant to the truth or falsity of a conclusion, or the validity of an argument.

That isn't to say, though, that recognition of bias, or the increased potential for bias arising from circumstances, is not useful. It does play a role in "critical thinking", because it serves as a flag for taking a more critical look at the statements to spot flaws in the facts or the reasoning. But relying on those circumstances to discount what he has to say is itself a flaw in reasoning.

Ipse dixit is a different issue: appeal to authority.

If it's slow, it's slow.

[matgorb]

Believe me, I'm 100% for Open Source, but lets cut the bullshit, if an application is even only 5 times slower to do essentially the same task (which means exactly what it means) then I said you can't start blaming it on how the files from one are better compressed or the internal is in german or other things, it is just 5 times slower whatever you said. I'm not saying these are not valid explanation, I'm saying it is counter productive. Yes it's free, but that's not an excuse, Open Source is also about doing better software, not only cheaper ones!

I would actually buy Office

[TWX]

I don't like Microsoft. I don't like Windows. I do, however, like Office. It's been a good office suite for a very long time. It's been very easy to use since I first started playing with Office 4.2. If Microsoft would actually release a version of MS-Office for Linux then I would probably purchase it.

Before everyone starts ranting about how this isn't good for GPL, or how I'm being bad by saying this, remember, the point of the GNU OS is for application developers to have a level playing field. Microsoft, like any other consumer software maker would be just as correct to participte in that kind of market as anyone else.

I use Open Office, but I don't agree that it's the best productivity suite. It is the best free productivity suite for Linux at the moment. Since Microsoft's product will always cost money, Open Office undoubtedly will remain the best free productivity suite; it will serve as a baseline. If vendors wish to make a commercial product that is better than Open Office and charge for that product it's their right to do so.

Re:I would actually buy Office

[Anonymous Coward]

No, the point of the GNU OS is to have a free (as in freedom) operating system, NOT so that developers of proprietary applications can have some advantage.

Re:I would actually buy Office

[kiehlster]

"Microsoft, like any other consumer software maker would be just as correct to participte in that kind of market as anyone else." I'd disagree. Microsoft may have a good product going with Office, but to release it on Linux would bring market loss to them. If you give the corporate world an office suite on linux that is powerful, rich, not bloated, standards complaint any every other bell and whistle, they'll all buy into Linux. What happens then? Microsoft's windows department goes bankrupt because most of their income is based off of corporate investments. Since Microsoft has all the anti-trust stuff to deal with, they've come down to splitting their business into sections. The Office group scratches the backs of the Windows group, and vice versus. Clearly it'd be a bad move for the capitalistic minds of Microsoft to put out Office on Linux.

Re:I would actually buy Office

[dnoyeb]

Office has been a great product. But after office 97 its been down hill ever since. Essentially Office is suffering from the same problem that doomed WordPerfect. Windows.

WordPerfect fell when it started, naturally, to use windows components for things the WP engineers had done a much better job at. WYSIWYG, windowing system, fonts, etc.

Office has become increasingly componentized. And even drag and drop was better in office 97. I can't even paste from one office product to another without using 'paste special' else it looks whack.

This is also the reason why office looks like its small on memory use. Because its componentized. Heck the JET database engine Access uses is part of windows even if you don't have office. So many other parts are like that. Thats also part of why it starts so fast.

Re:I would actually buy Office

[timcharper]

you can run microsoft office under linux - there's a tool called crossover office from codeweavers.org. You have to pay $40 for it, but it works really great!

You could also try and set up wine (crossover office is built on wine), but I couldn't figure it out... so I just paid for crossover office.

Re:I would actually buy Office

[canajin56]

That's why they were very careful to say "Private" memory. It wouldn't look so good if they said that the other 292MB (Number made up) used by excel is shared between all office applications ;)

Re:I would actually buy Office

[Anonymous Coward]

My first introduction to Word was Office 97. If I had a dime for every paper I lost because the file got corrupted, well, I'd have about $0.50. It had this nice "feature" where if you copied a file while it was open, it would be corrupted. So, if you were editing a file and you saved a copy on a floppy and then tried to save a copy on a hard drive, both files would be useless as would the original. You had to close the file and then either manually copy it or open it in Word and save it somewhere else without editing it first. How convenient. It took them until Office 2000 to iron out this little issue.

Yes, Word had a lot of nice features early on like continuous spell checking, grammar checking (mostly useless, but it would occasionally catch a legitimate error), and document templates. But it had at least as many bugs and quirks that shouldn't have existed in software that cost hundreds of dollars.

Re:I would actually buy Office

[legojenn]

Wasn't the office licence amended to allow users to only run office on Windows, not through an emulator?

Re:I would actually buy Office

[arivanov]

Close, but no cigar. OpenOffice worse component is the bloody Calc. It uses 4 times or so more memory then gnumeric for the same spreadsheets and eats considerably more resources. It is in fact the worst component in OpenOffice. This along with lack of decent chart/vector/diagram integration into both writer and impress are the main factors that hinder the adoption of openoffice in business. Writer is clearly better then MSWord and is getting better with every release. Calc has been going down since StarOffice 3.3. I have not really played with 4, but 5 was clearly worse then 3.3 and it has not improved since.

Re:I would actually buy Office

[falconwolf]

Microsoft may have a good product going with Office, but to release it on Linux would bring market loss to them.

While MS would loose some of the OS market they would more than make up for it in the Office arena. I've heard a few people say the only reason they stay with Windows is because of Office, well there's Office 2004 for Mac and I've also heard some Linux/Open Source users say they'd get Office as well if it were available. Actually if there weren't a market for it then CrossOver OfficeTM 5.0 Featuring Microsoft Office 2003 wouldn't of been made. I don't have the data to support this but I've heard that Office is the top income producer for MS, not Windows, and if they released it for Linux then they'd sale even more.

Falcon

Re:I would actually buy Office

[timcharper]

hmm, i've never heard that. I don't believe so... can you show some evidence?

Re:I would actually buy Office

[Anonymous Coward]

"MS would loose"

Fly, be free! LOLOLOL!!!1!!!111one

Re:I would actually buy Office

[Farroos]

I do, however, like Office. It's been a good office suite for a very long time.

Just like saying: I like apples. They were good apples for a very long time.

Re:I would actually buy Office

[Markus_UW]

Plus WINE Is Not an Emulator.

Re:I would actually buy Office

[terjeber]

While MS would loose some of the OS market they would more than make up for it in the Office arena

Nope, they wouldn't. In order to run Office you have to have Windows. People will invest in Office (noone invests in an operating system as such), so for each person who have a requirement to run Office, MS will sell one license of Windows. There are close to zero real alternatives to Office out there, so MS can not expand it's market share there. Porting Office to Linux will therefore increase sales with "exactly" $0. Each extra Linux Office sale will mean one less Windows (or perhaps Mac) Office sale. No gain for MS. In addition, they will find that their Windows sales dwindle due to less OS license sales.

Porting Office to Linux until there is a real competitive alternative would be insane for Microsoft. Once there is a competitive alternative, they should however, so as not to lose Office sales. Oh, and no, I have tried Open Office, and it is not yet an alternative.

Re:I would actually buy Office

[Taevin]

I just sacrificed part of my soul to read through the entire Microsoft Office EULA so I hope you're happy:

Installation and use. You may: (a) install and use a copy of the Software on one personal computer or other device; and (b) install an additional copy of the Software on a second, portable device for the exclusive use of the primary user of the first copy of the Software.

No mention in the EULA of requiring the user to install and use the application only on Windows. I could be wrong but wouldn't that put them in a precarious legal position given the fact that Microsoft is a convicted monopolist?

Re:I would actually buy Office

[MechaStreisand]

No no no, they changed it recently. Now it stands for WINE Is aN Emulator.

Re:I would actually buy Office

[Anonymous Coward]

No mention in the EULA of requiring the user to install and use the application only on Windows. I could be wrong but wouldn't that put them in a precarious legal position given the fact that Microsoft is a convicted monopolist?

It would also leave Mac OS users reeling, since after paying $100 for the Mac OS X version of MS Office, they read the EULA and find out they're not allowed to run it.

Re:I would actually buy Office

[falconwolf]

Porting Office to Linux until there is a real competitive alternative would be insane for Microsoft. Once there is a competitive alternative, they should however, so as not to lose Office sales. Oh, and no, I have tried Open Office, and it is not yet an alternative.

There is a market for Office on Linux otherwise Crossover Office wouldn't exist. Fact is is that CodeWeavers is providing software people are willing to buy. Others are using WINE to run Office in Linux as well. There are people using Linux now that would buy Office.

Each extra Linux Office sale will mean one less Windows (or perhaps Mac) Office sale. No gain for MS. In addition, they will find that their Windows sales dwindle due to less OS license sales.

I just checked online for the prices of Office vs Windows and the cheapest for Office 2003 as an Upgrade is $240, well actually $150 for the Student and Teacher Edition but the cheapest full edition was $400. For Windows the most excpensive was for Microsoft Windows XP Professional w/SP2 and it cost $300 whereas an upgrade cost as little as $100. Now if someone were to switch to Linux but still bought Office that's $400, or if they buy a new computer it very well could come with Office and I doubt the OEM paid half that, maybe for both Windows and for Office. Quite simply if someone were to switch to Linux and MS released a version of Office for it MS might lose a couple of hundred by not selling Windows but there are a lot of people who already use Linux who woud hand over $400 to MS for Office. MS would come out ahead.

I have tried Open Office, and it is not yet an alternative.

I've downloaded OO v2 but haven't installed it yet, which they need to make better! I don't know if I ever will, at least not on this computer. In two or three months I plan on getting a Mac Powerbook and will get iWork with it. Simply I don't see the need to pay $400 for Office, especially when all I may use much is Word.

Falcon

Re:I would actually buy Office

[Torham]

the point of the GNU OS is for application developers to have a level playing field.

I thought it had something to do with freedom.

Re:I would actually buy Office

[poofyhairguy82]

Writer is clearly better then MSWord and is getting better with every release.

Until OpenOffice's Writer can check grammer and do many other things Word can do (title case or autocomplete anyone?) its not comparible.

Lucky for me, Abiword is quickly catching up to Word (grammer checking in newest version) and will pass it this year I bet.

Re:I would actually buy Office

[jtev]

That was for a the developer's edition, ment for making office based applications. the licence said you could only make applications meant to be deployed on windows.

Re:I would actually buy Office

[arivanov]

FYI: Writer has autocomplete. In factr, it works better then the MS Word one as it has a better context adaptation. In MS Word I have to turn it off when I write technical documentation as it keeps coming up with phenomenally cretinous MBA/PHB style suggestions. In OpenOffice I do not. The suggestions are dependent on the text of what you have written so far and get better as you go.
As far as the grammar checker - fair point. But it is a much less hindrance then the lack of vector graphic (visio and friends) import. No matter how much I hate the stuff when I have to do trivial network documentation or presentations I have to power up the b*** MS sh***e.

Re:I would actually buy Office

[terjeber]

There is a market for Office on Linux otherwise Crossover Office wouldn't exist ... people using Linux now that would buy Office

There is a market for a lot of things, but that doesn't mean that there is a compelling market for Microsoft for these things. A small company can make lots of money with 100, 200, 1000 customers. No problem. If Microsoft adds a consumer product that picks up 1000 customers, it will be a huge loss for them. In that way there is no market whatsoever for Microsoft Office on Linux even though there is a market for Crossover Office. A customer brings cost, not only revenue. Microsoft will need a critical mass of customers before it pays, or, as I pointed out above, the real risk of losing customers to a competitive alternative. Neither is the case at this stage.

Simply I don't see the need to pay $400 for Office, especially when all I may use much is Word.

The other thing is that Microsoft doesn't much care abou this, the reason is simple, there aren't all that many like you buying Office. People either get it bundled or through their employer who's bought site licenses.

What do you expect??

[Anonymous Coward]

Just one question: What do you expect for paying NOTHING? Oh no! It takes a couple more minutes to load! Yup, that little feature is worth running out and paying $300.00 on MS Office. Put it all in perspective, people. Open Office is FREEWARE. And, as such, I find it an excellent value.

who needs it?

[tomcres]

heck.. who even needs OpenOffice, when you have ex!

Stop Whining and use the Source

[Fujisawa Sensei]

Have any of the people whining about code bloat/performance actually bothered to run a code pofiler on OO?

If you don't like it, use the source and present a solution.

Ok, here are my benchmark results:

[jbeaupre]

I used the data you suggested. My methodology and note-taking aren't perfect, but here's the lowdown

On a 3.06 GHz Intel with hyperthreading on (though it makes no difference) and 2 Gig of ram
OO 1.9.122 vs Excel 2002 SP3
3.4 Meg SVX file vs 191 meg xlm

File load: SXC OO ~3 minutes , XML Excel ~ 1 minute (not including time to unzip and open from withing excel).
Memory use (meg): min/typical/max OO 13/115/212 excel 4/45/65 (yes, 4 meg with a 191 meg file open, go figure !?!?!)
proc load: 100% during load times for both on "1" processor (HT did not help)

I saved the sxc file as ods and xls versions (hadn't figured out loading the zip into excel at that point). 3.9 meg and 49.5 meg file size repectively.

File load: ods OO ~1 minute, xls Excel ~3 seconds
Memory use (meg): min/typical/max OO ?/72/72 Excel ?/91/91
proc load: 100% during load times for both on "1" processor (HT did not help)

Seems the Excel has some advantage in extreme situations but the caveat mileage may vary seems to apply.

If OO would run twice as fast or use half the memory, I'd be willing to pay twice as much!

Re:Ok, here are my benchmark results:

[SatanicPuppy]

Excel and Access both only open a tiny amount of data at a time, which explains your memory usage...Excel is pretty good at loading more rows on the fly, unless more than a few columns are dynamic which slows the whole mess down, but Access is such that, if you have a table with 200,000+ rows, and you hit the "Last record" button, you can go get a soda and flirt with the girl down the hall for 20 minutes, because your computer ain't going to be doing anything else for a while.

Looking at the problem

[Great_Geek]

Instead of just talking, I looked at the data files. Several points are clear:

• OpenOffice loads reasonably fast, so it is not bloated in the classical sense.

• The example is 3 megabytes in .sxc and 188 meg in .xml. Interestingly, the file properties say "Patrick W. Jones" modifed Sept 13, 2005.

• OpenOffice 2.0 takes a long time to do anything with this example - open, write, anything.

• The example is basically a table of 16K rows by 13 columns of data, repeated for several sheets. There is no calcalation that I saw (in a very quick look).

• The GUI repaints very slowly on this example. Unclear if this is due to Java, more likely reason is that this is another symptom of the same slowness with large sheets

• This example may be unusual, but is not completely artifical.

I think it is clear that for large sheets (at least for this example), OpenOffice 2.0 calc is unusable. I am gusssing that there is a quardratic data-structure/code somewhere (without knowing anything about the source code, I am willing to bet that each time a row is needed, all the rows are scanned; or something like that).

On the other hand, this is a .0 release, so let's wait for the fixes.

What if???

[treo1167]

This may be an unusual concept, but what if MS's Huge Team of Developers have actually put out a refined copy of office, and OO being open source, hasnt reached that level yet?

This is a RAM issue period.

[billycub]

His two complaints are one and the same: that OpenOffice uses more RAM than Excel. I have 2 Gb of ram in my machine, and his 16-sheet file loaded in one minute flat, not three as his benchmark showed. This is within 25% of the 47-second time that zipped MS XML loaded in Excel, which should be the most useful comparison since OpenOffice uses zipped XML as well. So, it doesn't seem there is a huge processing difference between the two engines; the difference is much more likely that he was swapping memory out to disk. My conclusion: Fix the RAM problem in Open Office and this issue evaporates.

The original's still the best!

[tomcres]

The Mac version of Office is IMHO quite superior to the Windows version. It just works and it is much more unobtrusive. My main gripes with the Windows version have been it offering a little too much help when I just want to get some work done and the sometimes seemingly random manner that it shows special toolbars. Never had any of these problems in the Mac version. I hope someday they port it to Windows, even! ;)

Re:The original's still the best!

[astrodud]

I dont know what mac world you live in, but for me, office on my OSX box has been slow, bloated, and buggy (crashing every other time I use it). It has basically driven me to use NeoOffice which (while based on OpenOffice 1.1.4) is not much slower or bloated, but crashes less.

George Ou at work

[Anonymous Coward]

This article is so biased and misleading it's not funny anymore. That's George Ou for you, bashing open source software again. It is sickening.

erm..

[Turn-X Alphonse]

I kinda have to ask since this makes no sense to me.

How the hell is 1 second load time making the news? I mean really.. it's 1 second, over a week if you open OO every work day thats all of 5 seconds. So what a minute a year is going to some how make the world come crashing in on it's self?

Re:erm..

[AlXtreme]

As Mr. Jobs has stated in the past, you have to realize that millions of people are going to use your product. Suppose 10 million people use OOo, a minute a year would result to 19 human-years lost. For merely loading documents.

From my personal experience, OOo takes at least 20 seconds to load on a reasonable machine. Suppose I start OOo once every day (not very unrealistic) that would mean some 120 minutes lost. That would result in 2280 years lost over the population of OOo users. Tends to put load-times in perspective, doesn't it?

Abiword and Gnumeric do fine for me personally and start in a flash. Now if my boss would stop sending me .ppt's I could finally get rid of OOo for good. Speaking of which, the latest Abiword can handle ODF! \o/

Re:erm..

[Turn-X Alphonse]

but 1-2 seconds is nothing (20 is something). You probably waste more time yawning or sneezing in a day then you do in a year of 1 extra second loading.

disgusting programming

[tomcres]

That's truly awful design for this kind of application. I suppose you're going to tell us next that's it's also statically linked!

How much does OpenOffice cost?

[Anonymous Coward]

I can handle a few microseconds of delay or extra memory usage in return for a free software product that's a great alternative to MS Office.

Besides, Microsoft is evil.

Re:How much does OpenOffice cost?

[Anonymous Coward]

Three minutes to open a document is "a few microseconds"?

If you had a large, real-world file that took 3 minutes to open it every time you needed to see or edit it, that could easily end up costing your company more than just buying office!

Secondly, if you need to double the ram in any machine that is going to use OO.o, suddenly the software doesn't sound too free.

Free CRAP is not better more than an expensive but usable tool.

Re:How much does OpenOffice cost?

[Anonymous Coward]

whaddya you mean crap? its open source! therefore it is automatically better than evil M$ software!!

Re:How much does OpenOffice cost?

[Anonymous Coward]

Sounds like you're an Anonymous Microsoft Coward.

P.S. If it takes your computer three minutes to open *any* document, you need a new computer.

Fool.

Re:shoe on other foot this time? weird.

[Lawrence_Bird]

The bottom line is that just about *all* software that comes out these days is
bloated.  Perhaps someone who is an expert (or at least codes for a living instead
of personal use) can comment better - is it just bad assumptions? Bad technique? Bloated libraries?

As an example, and not to pick on java, I use a standalone java application
daily. The first version was released about four years ago had a working set
about 8Mb and a virtual size of about 50Mb. The latest version comes in at about 60
and 250.  Yes there have been enhancements, but the primary function remains the
display of real time data and is little changed.

Further, my daily running desktop applications, which have been more or less the
same for 7 years, now requires a commit charge under XP of 850MB and often near 1.2Gb (glad I have 2GB. Five years ago the same (under win2k) would have been 300-400Mb.  I can't point to any great increase in features or usability to warrant this increase use of resources.

Yes, what an improvement!

[tomcres]

If you run OOo 1.1.4 and OOo 2.0 side by side on windows, the 2.0 version uses about 10MB less memory when both have nothing open

I'm soooo glad that OOo 2.0 only uses 211 MB of RAM as opposed to 221 megs for 1.1.14. Now I can get rid of that swap partition...

Re:Yes, what an improvement!

[mikefe]

How many languages do you write in?

Now remove the dictionaries you have installed for the other languages, and your memory usage will go down drastically.

How much processing power do you need???

[kermitthefrog917]

In this age where the average new computer is in the 2.6 ghz range (im talking budget as well) and RAM is at a minimum 256MB (but is usually around 512), What is the average person gonna use all that power for? Email doesn't take much. Playing Freecell doesn't either. If I were to go to a normal office building, and see the average load, I'm guessing it would not be very high.

Now... As Computers get faster, and an Office Suite begins to provide more features than the average user even would think about using, what is the drawback to OpenOffice? yes, I know ooCalc2 sucks when it comes to scientific graphs and stuff. But does the average user ever do that? No!

I personally believe that for the general population, Processor speed and updated software is reaching a plateau. I mean, what else do you want in a word processor? Combine this with a massive amount of spare CPU cycles, and people are going to see that shelling out a couple hundred bucks for features that they won't use is ridiculous, and a reason against it definately will not be that they are concerned with their RAM having a quarter dedicated to their office suite. It does what it needs to do, and it does that for free. That's all that will matter.

Productive Time

[Kaenneth]

Far less employee time is wasted waiting for Open Office to run than is wasted on Freecell

It's C++ AND Java

[Anonymous Coward]

[root@lx OOo_2.0.0rc3_src]# find . -name '*.c' -type f | wc
        271 271 7711
[root@x OOo_2.0.0rc3_src]# find . -name '*.cxx' -type f | wc
    10802 10802 471294
[root@x OOo_2.0.0rc3_src]# find . -name '*.java' -type f | wc
      3250 3250 196385

The Java appears to be what slows things down but if you point that out you seem to get modded down around here.

I don't think Java fans can handle the truth.

Good questions

[jbeaupre]

I converted sxc to xls (see other post). Excel opened it in ~ 3 seconds. OO 1.9.122 opened in ~ 20 seconds, but used 137 meg. Would have opened in half the time except for all the auto row height stuff that it did. I'm beginning to suspect Excel is optimized in 2 ways: code size & IO speed. I suspect the IO speed trick is to only load what it required since the HD is busy whenever I go to another part of the document.

Has anyone profiled OOo

[D3m3rz3l]

I'm not sure how feasible it is to profile such a large program, but I'm sure Microsoft profiles the daylights out of their stuff. Do OOo developers profile things like the start-up time? After all, you can't start optimizing things unless you figure out exactly what is slowing it down. Is it the Java run-time engine? Is it because it needs to load a lot of libraries that MS Office does not need to (because of dynamic linking to Microsoft DLLs). Maybe when loading certain data sets, the program goes into a pathalogical state, creating hundreds of thousands of small objects? I don't know.

But things like analyzing profiling data and then optimizing are not fun to most people. Even more so if it means that an algorithm needs to be re-written. After all, if the "open file" operation needs a complete re-think + re-write, who's going to do it? It's not "fun". After all, the "open file" operation already exists. Generally, I think programmers like to build *new* things as opposed to fixing old things. And in this case, it's not even a matter of "fixing". It's a matter of rewriting. I presume that at Microsoft, if Word's "open file" operation (run with me on this for a minute) is uber-slow, then somebody is going to *have* to fix it, or not get a good performance review/etc. However, in the case of OOo if no one makes it faster, well, it does not negatively affect the person who wrote the slow version in the first place (not to discredit OOo authors or anything. They've done a phenomenal job given that they do this for fun and not profit).

Of course, there are an equal number of programmers who like to fix security holes and so forth, but patching a security hole is one thing, while re-writing major algorithms in a large program is another. There are of course some programmers who love optimizing code (Michael Abrash?). But I think they are far and few between. Very often, once something works, an attitude sets in that "It's working. Now don't break it". And optimization in it's early stages will often break things.

Re:Has anyone profiled OOo

[Anonymous Coward]

I have taken great pride in turning matlab algorithms written by professors that take hours to run into mostly efficient (and sometimes clearer) algorithms that take minutes or seconds to run.

although scientific programming is not the same as application programming (I only have to worry about a few hundred lines of high level code) I imagine that the "accomplishment high" would be just as great from fine tuning an application as a numerical model.

Re:Has anyone profiled OOo

[Anonymous Coward]

The problem is that optimizing something like this doesn't look like optimizing an algorithm. It looks like redesigning the engine.

For example, let's say that OOo loads a large file slowly because it has to parse the whole thing into in-memory data structures and MSO loads the same file quickly because it can load only the parts that are needed in order to display what's on the screen. MSO was always designed this way because it was originally written for systems with 1MB of RAM and no virtual memory. However OOo was probably designed ground-up on systems with virtual memory and lots of RAM relative to typical file sizes, so this sort of design never occurred to them. The whole engine will likely have to be rewritten from scratch in order to only load in the parts of the file as needed.

That sort of thing isn't fun like optimizing an algorithm. It's long, painful, and tedious, like refactoring.

dom

Re:Has anyone profiled OOo

[zzyp]

I disagree about the "programmers who love optimizing code are few and far between".

OpenBSD folks are fanatical about security AND optimizing. I know NetBSD also tries to reduce code bloat, I remember reading a comment where they proclaimed that they had added new features while cutting overall lines of code.

Re:Has anyone profiled OOo

[BasilBrush]

Don't assume everyone's likes and dislikes are the same as your own. Different people get satisfaction from working on different things. I happen to enjoy optimising slow code and refactoring ugly code more than pretty well any other coding tasks. And plenty of others enjoy optimising too, which is why so many people optimise too early - it's fun!

Re:Has anyone profiled OOo

[AaronLawrence]

Actually lots of programmers like to rewrite code, but it's not always true that the new code is significantly better :)

I imagine plenty of people have tried, on OOo and other projects, to rewrite big chunks, only to find that it's really hard to get it all working properly again, or that the result is actually worse, or all those seemingly simple patterns didn't quite work.

After all there are plenty of enthusiasts willing to hack into things, but not many people willing to do the 80% grind to get it back to the reliability it was before the rewrite.

MFC71.DLL

[AndroidCat]

No MFC? How strange. What is that huge MFC71.DLL file doing tucked into the Office directory? (A major pain for VS2003 MFC apps that can no longer assume that Microsoft has tossed the latest MFC DLL into system32 where all apps can use it.)

If Office uses zero MFC, then it should be safe to delete that file, right?

Re:MFC71.DLL

[bheer]

Odd, I _don't_ have MFC*.DLL in my Office11 folder, nor in Common Files/Microsoft Shared/Offic11 (XPSP2). But that could be because I don't have Outlook installed (and this guy who apparently was an Office dev says Outlook uses MFC). I'm not familiar with Outlook, but I'm pretty sure of my facts about Excel and Word.

Re:MFC71.DLL

[AndroidCat]

That's possible. This was a Dell box with all of Office pre-installed.

It doesn't surprise me that MS doesn't use MFC for most of Office. It's obvious to anyone forced to use MFC that Microsoft hasn't eaten their own dog-food in quite some time.

Re:MFC71.DLL

[bheer]

It doesn't surprise me that MS doesn't use MFC for most of Office

The point is, when Word and Excel were written MFC didn't exist. Heck, 'C++' was probably a dinky little preprocessor from AT&T. Like LordSah said, MS would very rarely rewrite giant chunks of UI just to support a new technology (the one time they probably did was when they transitioned their apps from DOS->Windows).

It's obvious to anyone forced to use MFC that Microsoft hasn't eaten their own dog-food in quite some time

MFC isn't as bad as some people make it out to be. But in reality its not the sort of thing a products shop like MS would be using. MFC is all right for QnD C++ GUIs (inasmuch anything in C++ is QnD), but if you want to do anything serious, like write a vector editing design surface, you have to roll so much of your own code that it actually makes sense to not use MFC at all, instead develop your own framework.

That said, MFC's meant to be a thin OO shell over Win32 with sensible defaults for small apps, and it's good at what it does. The problem begins when people assume that CScrollView should be able to efficiently deal with their 30MB vector diagram document, or that CArchive should efficiently handle their boardgame state.

Finally, MFC is *old*. It doesn't use smart pointers (AFAIK) or templates, for example. Anyone programming in Win32 and C++ would probably do well to use ATL (or ATL-like libraries) and things like Boost.

For people looking to MFC-style RAD, the answer is very clearly .NET. C# and friends run rings around anything C++ can ever offer in the 'user friendly' department. But that's all right because I use C++ for control and efficiency, not RAD.

What ?! Open Office bloated ?!

[beforewisdom]

What ?! Open Office bloated ?!

This belongs in the "is this news?" category.

Many people wrongfully take it for granted that Open Source software is always better on the back end.

Who's bloated and where?

[shotfeel]

I was wondering how much of the RAM footprint difference was due to Office relying on Windows code. So just for the fun of it I fired up Excel on my Mac. 22.94 MB of real memory being used for Excel, 34.14 for Word. Compare that with 7.10 and 9.81 for Excel and Word on Windows and 37.54 and 37.66 for Calc and Write on Windows. Anyone running OpenOffice on a Mac want to add another data point where MS doesn't have code "hidden" in the OS?

Re:Who's bloated and where?

[shotfeel]

Just to clarify, the numbers for Office on the Mac are from my machine (OS X, Word X for Mac, Exel X for Mac). The Windows numbers are all from TFA.

Re:Who's bloated and where?

[asdfgl]

In what way does Office have code "hidden" in the OS?

If office suites didn't use code included with the OS, then they truly would be bloated. You just can't compare code size the way you did (of course you can, it's just really hard to come up with a valid conclusion that's all :-)).

Re:Who's bloated and where?

[Anonymous Coward]

Mac Office isn't really a straight port though, so a direct comparison is difficult to make. And if it were a straight port the Mac-version would probably use some extra UI-library the Windows version doesn't need (eg. some MFC-on-Cocoa layer), like some Linux ports have used.

Re:Who's bloated and where?

[ThaFooz]

Anyone running OpenOffice on a Mac want to add another data point where MS doesn't have code "hidden" in the OS?

I've wondered that myself, but unfortunatley its rather difficult to make that comparison. MSO for the Mac uses Apple's native Quartz windowing system, whereas there really isn't a full port of OOo to the Mac yet - you have to choose between OOo for X11 or NeoOffice (Java-heavy OOo)... both of which tend to be incomplete and/or several versions behind. Since the X11 emulator and JVM are launched on demand of the apps, OOo will always feel quite sluggish, and its difficult do determine how much of the RAM footprint is due to that key difference. Its a real drag, which is why I abandoned OOo in favor of MSO on my powerbook, whereas OOo on my Linux & Windows PC's is just fine and dandy.

Kind of a tangent, but I would be interested to see the memory comparison of MSO to iWork. Of course, then we're trusting that Apple doesn't "hide" any code ;)

Re:Who's bloated and where?

[mrchaotica]

The difference is that MS Office probably uses standard Windows APIs, whereas OpenOffice has to use the Java API and/or implement everything itself for cross-platform compatibility. So yeah, "If office suites didn't use code included with the OS, then they truly would be bloated." is probably more accurate than you thought when you wrote it.

Re:Who's bloated and where?

[shotfeel]

I would be interested to see the memory comparison of MSO to iWork

Just fired up Pages (the iWork word processor). It runs at 18.31 MB an the same system I used fo the Mac Word and Excel measurements.

Re:Who's bloated and where?

[microtoph]

... and that's exactly the point, where, as a Windows user, I have to prefer an office suite that's written for Windows specifically. What do I care whether it runs on MacOS or Linux, too? I only use one OS at a time, so I'd rather have something that performs well on that OS, than wasting my time with a bloated implementation, just because, theoretically, I could use it on another OS. With the huge prevalence of Windows OS's, there should be an OSS office suite taylored for Windows -- in the marketplace, outperforming other Windows office suites will be much more interesting than being multi-OS.

Re:Who's bloated and where?

[dave420]

Office uses lots of features already present in Windows, so it doesn't need to be running them itself. It can simply use the OS that it's in. If you take office out of Windows, and put it in another OS that doesn't have those features, then it has to bring them to the game itself, hence your perceived higher memory use.

Re:Who's bloated and where?

[alerante]

NeoOffice/J on my machine with a blank Writer document uses 25.12 MB of real memory and 1.06 GB (!) of virtual. (NeoOffice/J is based on the 1.1 codebase, but I'd imagine a lot of Mac users would go for it because it integrates better with the OS X environment.)

Re:Who's bloated and where?

[GWBasic]

When programming to a GUI, one holds many structures in the program's memory space that are defined by the operating system. Windows and MacOS use very different GUI APIs, and thus I find it difficult to accept your commant as an "apples to apples" comparison. (Heck, I remember that Doom 1 required twice as much RAM to run it on OS/2, even though it was the same binary that DOS used!) If Windows requires that a program allocate 500K for a window, but MacOS requires 1 meg for a window, then all MacOS programs will use more memory.

Re:Who's bloated and where?

[mykdavies]

Good point - I've just compared Office/X with NeoOffice/J (perhaps not really fair on OOo, as NeoOffice brings Java into the equation as well).

(All on a 1GHz Powerbook, 750MB RAM)

Footprint with a blank doc:
NeoOffice/J Writer 74MB
MS Word 20 MB

Footprint after loading a 750kB html file:
Writer - 103MB
Word - 31MB

Launch time:
Word 17 seconds first time, second time was 6 seconds
Writer 47 seconds first time, second time was 17 seconds

Word Document open time:
Word - 3 seconds
Writer - 16 seconds

Opening a .txt file (750kB)
Word - 3 seconds
Writer - 5 seconds

Opening a .html file (730 kB)
Word - 5s
Writer - 5s

Not looking too good for OOo there...

Re:Who's bloated and where?

[ciw42]

It's a fair enough point, but when you're actually using the software, you'll probably find very little difference in terms of speed, and more importantly, productivity. It comes down to personal preference and past experience more than anything else, and software cost/cross platform capability are not going to be an issue for everyone. For those who feel it an issue, they're certainly not going to be upset wasting a few seconds a day when they know they've saved £250+ and can use the same software freely regardless of which OS they are swat in front of.

Let's say it takes OOo 15 seconds to load your first document of the day, and MS Office only takes 5 (which based on my experiences are about right) by the end of the working week you'll have saved yourself less than a minute. Big deal.

In practical terms, the fact that if I'm running OOo's Write and it has a 20Mb bigger memory footprint than MS Office's Word is of very little consequence for the average user running a machine with 128Mb or more of memory. Unless you've got half a dozen apps open at the same time, or you leave OOo idling in the background for long enough to let Windows swap it out to disk, you're not going to have a problem.

This whole news story smacks of FUD in the aftermath of the Mass. decision to only use open formats for their data and documents. Now, I wonder who could be behind such an effort?

Re:Who's bloated and where?

[k12linux]

in the marketplace, outperforming other Windows office suites will be much more interesting than being multi-OS.

Not if you happen to be someone who uses multiple OSes or an enterprise which uses multiple OSes. Also, if you are an enterprise considering saving $250,000/yr on MS licenses by not using Windows everywhere then you also probably care a great deal about cross-platform compatibility.

Re:Who's bloated and where?

[Sledgy]

This is very similar to the memory usage comparisons between Internet Explorer and Firefox. IE using native windows controls (and much of it's codebase intergrated into windows) and Firefox rendering it's own widgets, although K-Meleon (a browser that uses the gecko engine but with native windows controls) uses a simlar amount of memory as IE and loads just as fast.

Re:Who's bloated and where?

[mrchaotica]

It just depends on where your priorities lie. Personally, I care a lot more about using a non-proprietary file format than I do speed. I also care about cross-platform compatibility, because I regularly use all 3 major OSs (I prefer Mac and Linux, but am sometimes forced to use Windows). For example, even though Safari is in many ways a superior web browser, I use Firefox even on the Mac because it works the same on all platforms.

Re:Actually I've really seen the opposite

[spitzak]

OpenOffice is really a partially-commercial product, which may explain why it is the way it is. However for most OSS I feel the opposite of what you say is true. Programmers do *not* work on features, instead they work really hard on making the part that exists as fast and elegant as possible.

Now before you say this is good, remember that the unimplemented features includes things like documentation, installation and configuration programs.

Suspicion on size

[Todd Knarr]

My suspicion: Calc is holding a parsed XML document tree for the spreadsheet in memory, while Excel reduces it to an internal structure and recreates the XML when saving. Holding the document tree makes for more memory usage, but it allows for preserving of unknown XML without mangling it. Parsing to an internal structure is more efficient, but anything not recognized will be lost when saving.

NB: I don't think Calc works on the parsed XML, it probably has an internal representation to actually work on which is then tied back into the XML document so the XML can be updated to reflect the changes in the data.

No problem

[smchris]

"Wife! I NEED a new dual-processor multi-gigabyte machine. No, OF COURSE, it isn't just for games. How could you think even think that."

Nonetheless, I'm finding that I'm opening Abiword about 1/2 the time these days. I was a WordPerfect fanatic in the day but since Word set the standard for lowest common denominator for slapping simple text on the screen, I'm finding that Abiword fills those needs nicely a great deal of the time.

OOo stength is the the ability to improve

[Anonymous Coward]

It is alright if OpenOffice.org uses more memory or is slower. The strength of the product is the process of community involvement and the ability to change. This might focus the community on memory useage, which would be a good outcome from this article. OOo has other strengths, such as cross-platform, price, PDF export, no need to track licenses (very expensive process), etc.

My Test

[b3d]

I'm using Office 97 and OOo 2.0 on Win 2K. I'm using the OOo Quick starter, which means that soffice.bin is resident and taking about 15M. I'm also using MS Office Startup, which is soffice.exe taking about 1.1M. When I start Winword.exe it takes an additional 7.5M, and open a simple document that merely has the text "This is a test" it take another Meg or two, and only takes a second or two. Now when I open an OOo document (.odt), there is no other process that starts, but soffice.bin shoots up to 42M! However, it opens in two seconds.

So, while I do see more memory usage, a real speed test is not really very different. I suspect in the article, he is not using the OOo quick start, whereas he is using the Office Startup. And how much Office stuff is being hidden by it's hooks deep into the OS, and thus it's startup time is part of the OS startup.

This is not to say that OOo couldn't be improved. I'm sure it will be. But OOo 2.0 is an excellent tool to have in anyone's toolbox.

Peace,
Jim

Note: GP is correct

[shark72]

"You clearly don't know what an ad hominem attack is."

The GP does indeed appear to understand the subject. I think the confusion lies in the fact that there are various types of ad hominem attacks. In this case, this is what's known as a circumstantial ad hominem.

The wikipedia article explains this well. If you believe the wikipedia article to be incorrect, you may want to take the time to edit it.

"But when Ou, who has a long and easily verifiable history of writing articles that disparage open-source software, says the same thing, his words should be taken with a generous pinch of salt."

Ironically, you have made an ad hominem attack yourself. From the wikipedia article:

Ad hominem circumstantial involves pointing out that someone is in circumstances such that he is disposed to take a particular position. Essentially, circumstantial ad hominem constitutes an attack on the bias of a person. The reason that this is fallacious is that it simply does not make one's opponent's arguments, from a logical point of view, any less credible to point out that one's opponent is disposed to argue that way.

But I'm not surprised that you're incorrect, since Anonymous Cowards usually are. ;-)

Re:Note: GP is correct

[Anonymous Coward]

I frequently argue that C# is a better language than Java and get attacked for being a Microsoft shill. That's an ad hominem: there is no evidence that my technical preference for C# is related to any stake I have in Microsoft (in fact, I don't have a stake in Microsoft, and I strongly recommend you use Mono, not .NET).

If my history did suggest that I have a stake in Microsoft, then that would be a reasonable fact to point out and doing so would not constitute an ad hominem.

Re:Note: GP is correct

[Anonymous Coward]

But I'm not surprised that you're incorrect, since Anonymous Cowards usually are. ;-)

The cowards are the site operators. Registration is a crutch for the weak. I have explained this before...

Re:This wasn't tested on other platforms

[Anonymous Coward]

There should be much more to this test.
It should be dismissed as crap.

Problems with OpenOffice. Solution? Fork It.

[sweetnjguy29]

I am happy the OpenSource community is talking about what a horrid mess OpenOffice is. I like and use OpenOffice. But I have noticed that OpenOffice takes a LONG time to launch the application. And I have noticed that it takes a very long time to open documents. And OpenOffice tends to crash on occassion. On the other hand, it is free to use. So, like the article says, I have to choose between time and price. Shelling out $300 per computer is just not an option for me...so, there ya go.

I think the solution to the problem is the forking of the OpenOffice project. Personally, I use only the word processing and the spreadsheet programs. I think that most people only use these two parts of the program...so why not focus on these two? Make them the best. Make each program functional apart from the other. Make each of them lean and mean.

The problem with OpenOffice is that Sun is providing the code. Thus, the openness of OpenOffice is tied to Star Office.

As an end-user, I hope the developers of OpenOffice make it the industry leader, not the industry's Yugo.

Why isn't hardware vendors in love with OpenOffice

[Anonymous Coward]

This surely is the proof they need that OpenOffice is GOOD FOR HARDWARE business, as it compell users to upgrade to more memory, faster CPU(s,) bigger hard drives...

Where's the cost comparison?

[bigbigbison]

Microsoft office costs infinitly more to purchase than OpenOffice.org.

Re:Where's the cost comparison?

[narcc]

Microsoft office costs infinitly more to purchase than OpenOffice.org.

Funny, I thought it cost ~$400 more...

Re:Where's the cost comparison?

[bigbigbison]

I screwed up I meant to imply that it was infinitly times more expensive since zero time infinity is still zero. To say it is infinitly times more expensive probably isn't mathmatically accurate either...

Re:A problem for OOo

[symbolic]

After playing with the test files (only on OpenOffice mind you, since I don't run Microsoft software), I have to say that I'm a bit surprised by the performance. The XML file used by Excel is 147Mb in size, and yet winds up taking only 47MB of memory when loaded. I'm a little puzzled at how it can rip through 147MB (of XML no less) in only a few seconds though. With the huge memory requirements for OpenOffice, I can't help but wonder if it's storing the entire XML DOM in memory.

Whatever the case, the development team should really focus on this. Waiting 3 minutes to for a document to load (as well as saving to an alternate format) is excruciating. In fact, OO crashed while saving that test file to the ods format.

mmm, I disagree

[Praeluceo]

I say that Microsoft is about $500 too bloated. That extra 5 seconds on initial load, and 2 seconds to open a file are worth the half a grand it'd cost for me to purchase Microsoft Office 2003. Ehh, $499 for an office suite, I tell you, that's just insane. I'll stick with OpenOffice.org thank you ver much, it's more than capable. Happy business user of OpenOffice.org since 2002.

That's such 1990's thinking...

[sterno]

Technological advances render all but the most dramatic processing demands almost moot.

It used to be true that a given process would be run exponentially faster over time by the growing power of processors. So if you wrote bloated crappy code, within two years it worked fine because all the processors got better. The problem is that we're running into a wall as to how high they can clock the processors because of the heat and power requirements.

The solution has been to switch to a multi-core processor that runs at the same or even sometimes lower frequencies. This works great but it has one HUGE caveat: the code must be able to run in parallel. Code that is being written today, by and large, doesn't account for this. Sure there's threading and all that in much of today's code, but not quite such that you're seeing those same exponential increases in performance.

To write fast code today you have to be able to write code that can break down into numerous discrete chunks that can all work in parallel. With each new generation of processor, this is going to become more and more critical. Right now we have 2 cores, then we'll get 4, 8, etc. With each generation, you'll get faster performance, but with each generation, the code becomes more complex because it must be broken down into smaller pieces to take advantage of the extra cores.

Having said that, I suspect that the OO people are, at the moment, more focussed with creating functional parity between OO and Office. Sure, sometimes it will be noticebly slower, but if it can do everything Office can do, then they can refactor the innards down the line to improve performance.

Re:That's such 1990's thinking...

[jgrahn]

To write fast code today you have to be able to write code that can break down into numerous discrete chunks that can all work in parallel.

Or simply write code that uses the right data structures and the right algorithms. Most of us don't write ray tracers or encryption algorithms, but things that should not be CPU-bound on any machine made in this millennium ...

Quick ...

[Anonymous Coward]

Look at your "desktop".

What Java Apps are running?

What C/C++/C#/ObjectiveC apps are running?

(Whaps moron over the head).

Do you "get it" now?

XML should not be the standard for everything

[olddotter]

I agree with the parent post. XML's strengths are its near human readability. Which makes it easy to design or debug software that uses it to exchange data. But it is a wastefull format for storing information. XML is data bloat.

Don't let Ou's argument dissuade you from freedom.

[jbn-o]

This kind of analysis comes up from time to time. It will come up again. I'll take OpenOffice.org over the proprietary alternatives because OpenOffice.org is free software.

I'd rather have a slow free software office suite than a faster proprietary office suite. If I cared enough to fix this in OpenOffice.org, I could do it or hire someone to do it for me. But I'll never fully learn what the proprietary alternatives do when they run, I'll never be able to fix the things I don't like about the proprietary software, and I'll never be able to help my community by sharing improved versions of the proprietary software. I want fast programs but not at the expense of my software freedom.

When this issue is framed in the terms the open source movement uses, proprietors can often gain (or keep) a client. This movement doesn't object to proprietary software, it claims that when more programmers have access to the program's source code and are allowed to change it, the program may see some improvement. This philosophy is one which places business priorities first. The social and ethical effect on the public—the helplessness one feels with programs they can't inspect, change, or share—is reduced to "ideologial tub-thumping" by the Open Source Initiative, which started the open source movement and defines the term "open source". Free software, on the other hand, places a philosophical focus on caring about society, not chiefly business interests. The free software movement talks about giving me the freedom to decide how the program should work for me. This stresses placing the limitations of what I can do on me: I get to choose how much programming I want to learn and I get to decide someone else to do for me. And, most importantly, the free software movement celebrates the spirit of voluntary cooperation, what keeps society from being a dog-eat-dog jungle. I don't object to the open source movement's priorities in themselves, but I don't think they go far enough to help improve society, and I do think it is computer user's job to care about what kind of society we are allowed to have.

Re:Bloat? Don't talk to me about bloat...

[MROD]

Surely EMACS stands for Eventually Malloc's All Computer Store? :-)

If a file can't be created with cat(1) then it's not worth creating.

About MS Office 2004 for the Mac

[Jerry Smith]

We have a mixed environment, with mostly W2k servers and about 800 clients, half of them Macs (10.4). Users have a user-directory on a file-server, with a roaming profile.
First time a user logs in, and starts Word, Word plants over 80 megs of fonts into the users' profile.
All 2000 users... In groups of 20 or more.
That's 1.6 gig within 10 seconds... ever heard a server choke?
And we surely enjoy storing 160 gig of MS-fonts on our raids too.

Children, grow up and admit that OSS isn't perfect

[Theovon]

I'll begin by saying that I mostly use Linux, and I use OpenOffice even on Windows when I can help it. One reason is that I don't want to give money to Microsoft, but there are other reasons as well, including my belief that Free Software is the key to the advancement of IT in the future.

But this situation is pure hillarity. OSS fans have their list of reasons why Linux (or some Linux app) is better than Windows (or some Windows app). Two reasons near the top are that Windows is slower and more bloated. These reasons are sited often and are part of the OSS mantra.

So I find it incredibly ironic that now that the shoe is on the other foot, the tables are turned, etc., that these very same people are dismissing "bloated" and "slow" as unimportant.

No, you idiots. "Bloated" and "slow" are ALWAYS bad, even when they apply to an OSS application. That means there's something wrong with OpenOffice.org, and if you have half a brain in your head, you have to accept that it's broken for that reason. That doesn't mean you should stop using it or feel disillusioned. And defending your beliefs in the face of this embarrassment just makes you look stupid and inconsistent. HAVE SOME FREAKING STANDARDS, and have them ALL THE TIME, not just when they make your favorite thing look better. It's time for you to have egg on your face, admit it, and take it like an adult. And then the next thing you need to do is stop wasting your time and fix the problem.

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[jcole]

I was wondering how much of the RAM footprint difference was due to Office relying on Windows code.

Good question.

I use Open Office on Linux everyday at work and don't see it being as slow. Can we see some benchmarks of Open Office running on Linux/MacOS/FreeBSD/Solaris?

-Joe

Re:Bloat? Don't talk to me about bloat...

[BreadMan]

Emacs compiled for windows (20.7.1 i386-*-nt5.0.2195) uses 5,420K (says the control panel)

Not much bigger once it loads a trival C++ file and therefore loads cc-mode.

In fact...

[Dystopian Rebel]

What you are saying is reversal of proven tendencies.

In short, I say... Rubbish!

There are excellent open-source programs as well as bad ones. But I don't know very many that are full of bloat caused by unnecessary, unoptimized code.

Microsoft is no idyllic ship of disciplined software engineering. When was the last time you found a **flight simulator** inside Gnumeric? When was the last time you discovered a half-brained implementation of a dialog or complete feature in Monoposoft Word? (Clippy isn't the only instance of useless crap.)

I don't find OO to be intolerably slow. However, for those who do, someone else has pointed out that disabling Java improves the performance of OO significantly.

Whatever the foregoing arguments, the TREMENDOUS advantage of OO and most FOSS is something that Microsoft cannot TOUCH: Cross-Platform Functionality.

I edit OO documents at work, take 'em home, edit in OS X and FreeBSD, and take 'em to work the next day. This kind of flexibility and convenience outweighs the few virtues that Redmond can claim for MS Office.

It's open source!!!

[fok]

FIX IT!!!!!!

Yup

[planetfinder]

Well said. Defending bloat and sloth in this case sounds almost as hypocritical as Microsoft arguing for open standards in every instance where the proprietary standard isn't theirs while arguing that their proprietary standards are in the best interest of the consumer.

Re:shoe on other foot this time? weird.

[freeweed]

god help you if you want to add captions to your figures, or use "styles", or insert an equation, or do just about anything a good word processor should let you do

You know you're old when you parse a statement like this and the first thing that comes to mind is "figures? a good word processor should process words".

It still amazes me that we're using word processors to do desktop publishing. And get off my lawn!

Does prefetch help OOo too? If not, why not?

[Rob Y.]

I just took a look at the Prefecth folder on my WinXP box. It looks like there are links there for Acrobat and Firefox stuff along with a whole lot of other stuff.

Does this mean that prefetch happens automatically for frequently run apps? If so, why wouldn't OOo benefit from this as well? Is it not structured in a way that Prefetch can help it?

Tis true....

[jmorris42]

> nothing makes me aware of how much I need to upgrade my processor like starting OpenOffice.

That was the old deal between Microsoft and Intel. Microsoft would release software that would be ever more bloated and slow to drive sales of new hardware. But they haven't been keeping their end of the deal at Microsoft these last few years. XP was more bloated than 98/ME but not by enough to justify a 3+ GHz Processor. Of course ANYONE who makes a deal with Microsoft eventually gets shafted, it is just in their nature.

But still, Intel can't afford to overly annoy Microsoft even after they got shafted by them yet again by Microsoft picking IBM/Power for the Xbox 360 over Intel.

But I am suprised they aren't quietly pushing OO.o on exactly the ground you mention, that widespread adoption of it would be just the ticket to drive a round of hardware upgrades.

They should use Javolution xml parser.

[Anonymous Coward]

They should use Javolution open-source XML parsers (pull or sax parsers at http://javolution.org./ They are 3x-5x faster than any other Java xml parsers!

Re:Children, grow up and admit that OSS isn't perf

[jofi]

Excellent post, especially about skewing your standards when they make your favorite thing look better.

I have noticed another inconsistency that the same users you describe will deny a broken application being caused by Linux itself. But if any app breaks in Windows they are quick to judge Windows itself as being the problem, instead of their incompetence or something caused by another 3rd party app (or even faulty vendor device drivers), or maybe it is just the app is broken. EX: "OpenOffice is soooo slow F%(&%king Windows! All Bill Gates' fault!!11111oneone"

Excel is a special case...

[argent]

The Excel team at Microsoft have always been a bunch of hardcore performance-and-utility fanatics, and the quality of Excel reflects that. In my opinion it's the only component of Microsoft Office that's worth anything, and I dearly wish it was available as a separate program so fat incompetant slobs like Word could be left to scrounge for users on shareware sites.

To maintain performance and compatibility, they refused to get drawn into the COM morass for many years... they interoperated with but didn't depend on COM. At one point they were even using their own compiler. Setting OOO Calc up against Excel is like comparing a donkey to a thoroughbred, and never noticing that the rest of the horses in the stable with the thoroughbred are broken down old screws.

Re:Excel is a special case...

[chris_eineke]

I guess you could say that they excel at what they're doing. ;)

Re:Excel is a special case...

[Buelldozer]

Uhhhhh...

http://www.officemax.com/max/solutions/product/pro dBlock.jsp?BV_UseBVCookie=yes&expansionOID=-536907 354&prodBlockOID=537152275

It's been this way since about forever now, and I mean that. You've been able to buy Excel seperately since Excel was originally released!

Re:Excel is a special case...

[argent]

That just makes the sheep-like behaviour of the people who buy the whole of Office despite its complete lack of buddha-nature even more depressing.

Par for the course? (even "right"?)

[Bob9113]

To what extent is this just the proper natural evolution of a large scale application?

Step 1: Functional demo, very lacking in features and stability. This would be StarOffice up through the 5.x series, and the OpenOffice 0.x series.

Step 2: Dramatic increases in stability and completion of all the major technical functions, but with a somewhat clunky or non-intuitive interface. OpenOffice 1.x.

Step 3: More user friendly and natural interface, but performance is not yet up to par.

Step 4: Performance optimization.

Each step is the natural evolution from the prior state. The initial state is an idea, which leads to a functional demo. The functional demo gets poked at by a few outsiders who say, "This might be a good idea, but it doesn't support features X, Y, and Z, and it crashes all the time." That feedback leads to the incorporation of new features and advances in stability. Then a larger group of outsiders uses it and says, "Yeah, this is getting good - it does everything I need it to, but the interface is a little goofy, so I'm sticking with my current solution for now." That feedback leads to user interface improvements. Those improvements lead to a much larger group using the software, and more people using the software full-time, those people say, "Wow, this is really well done, but look at how much (CPU|RAM|disk space|bandwidth) it uses." Which should, inevitably, lead to performance optimization.

That sounds like the natural sequence to me. In fact, that whole process - release, listen, refactor, wait till the end to performance optimize - has always been a big part of successful projects and is now becoming a big part of standardized software development models like those that come under the Agile umbrella. It would be worse if there had been a lot of unnecessary performance optimization that had lead to an unmaintainable code base.

Re:Par for the course? (even "right"?)

[jdwyah]

bug 4914 "normal" view option needed http://www.openoffice.org/issues/show_bug.cgi?id=4 914 Once OO fixes this I don't care how much bloat they add.

Re:Par for the course? (even "right"?)

[Anonymous Coward]

My company (and I personally) would never allow a product to be stamped 1.0 until all of those steps had been completed. You map features to the next major version number, implement them, optimize, and then you can stamp it with the desired version number. OO may have all of the features that were mapped for 2.0, but until it performs well enough to be usable it should not be allowed out of the beta state.

Re:Par for the course? (even "right"?)

[David Off]

Step 1: Functional demo
Step 2: Dramatic increases in stability
Step 3: More user friendly and natural interface
Step 4: Performance optimization.
Step 5: ???
Step 6: Profits!

Re:Par for the course? (even "right"?)

[Kelson]

There's even been performance improvement in step 3.

I could hardly stand to start OpenOffice in the 1.0 days. If I needed to work on a formatted document or a spreadsheet, I tried to use AbiWord or Gnumeric whenever possible, using OpenOffice only when I needed the extra features or stability... and I tried not to close it. (Gnumeric's been rock-solid, but I've had lots of problems with AbiWord crashing, losing text or formatting, etc. No, I don't remember what version it was.)

More recent versions, and especially the 1.9 series when it was first included in Fedora Core, are *much* faster, and now I have no qualms about just firing up OOo.

Re:Par for the course? (even "right"?)

[Bob9113]

An A/C comment at level 0. I guess it is fruitless to respond, but what the heck.

My company (and I personally) would never allow a product to be stamped 1.0 until all of those steps had been completed.

So you're one of those VC sucking vaporware companies, eh? You can't do all of those steps without real world results. How, precisely, do you intend to do performance optimization without real world usage? Or are you saying that you optimize your software to suit your imaginary beliefs about how it will be used in the real world?

It's not a question of whether you do those steps before the first public release. It's a question of whether you compromise the end result with shoddy simulations of those steps in order to make the false claim that you have performed those steps before release.

Don't get me wrong, I'm not saying you shouldn't have solid fundamentals - you shouldn't use known-to-be-slow algorithms in heavily used sections of the code - but that's not performance optimization. Performance optimization is shifting the focus of a section of code from clarity to high performance. Doing that without knowing that the section in question is a problem is pure stupidity. Without real world usage data, you cannot know which sections are a problem.

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[trentblase]

Yeah, the article didn't mention how much ram the crappy "office startup assistant" takes. I think the real comparison is this: 1) Boot into windows with all unneccessary services off. Load up excel. Check total memory use. 2) Boot into linux with all unneccessary services off. Load up calc. Check total memory use.

Bloated?

[Anonymous Coward]

In the immortal words of Tucker McElroy:

OpenOffice, bloated? Suppose we ain't got no Word and go in there and start using OpenOffice anyway? Whatcha gonna do about that? You gonna stop us, Stein? Ha. You're gonna look pretty funny tryin' to eat corn on the cob with no fuckin' teeth.

Open Office Killed My Printer, and I Cried

[The Dark Caller]

With Java on, Open Office more often than not kills my poor PC, which is a relatively speedy AMD 1.7ghz with 512MB of RAM. . . to such an extent that it apparently killed my 6 year old Canon.

I only blame it on OO.o because it worked fine earlier in the day when I was printing w/ MSO, but the second I installed OO.o and printed, it went kaput. Le sigh. I miss ol' Canny, she was a good printer.

Seriously though, I don't have the time or the system resources to load OO.o anymore. P2P apps, chat apps, browsers (At least I use Firefox), WinAmp, WordWeb, and the occasional Warcraft 3 session don't need an oversized freeware app sucking their precious RAM. Unless OO.o can offer me comparable quality at a lower price point, I for one will make sure that my money is well spent.

You know, by switching to OSx and just pirating MSO :)

office is slow as dirt

[codepunk]

MS Office is slow as dirt on my linux box...see what is the point, come back when I can do
a speed test on linux.

Hello?!

[imsabbel]

Look at your timings. 30s real time, 3 seconds usertime.
Your first start was totaly HD-bound, while your 2nd start came from your diskcache.
That no magic, you know, and not even knowing such basics just seems like a BIG hint that you shouldnt even try to benchmark.

Re:Hello?!

[Shayde]

Jeez, tone down the abuse bits a few notches, eh?

You've answered one question - where the performance bump comes from on the second running. I'm all over that (this machine has 2gig of memory in it). But my point I'm trying to raise here is that there's so many factors that influence performance issues, that this fellow at ZDnet going "OO BLOATED! EXCEL GUT!" is really jumping, IMHO, to the wrong conclusion on the flimsiest of data.

One other little tidbit. From waht I'm understanding of the fellows tests, he's running OpenOffice under WINDOWS, and crowing that Excel takes only 3 seconds to load. One would point out that Excel has to load almost NOTHING to start, while OO needs to load it's entire environment to get started.

Oh, I tried to run Excel on my Linux box to compare results, but I can't seem to find a version available it. (:) for those without a sense of humor)

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[JonathanBoyd]

Isn't that a rather artificial situation? Ok, Microsoft Office may 'cheat' a bit by having some stuff built into Windows, but if that stuff is running when you use Open Office and the real world memory use for Office is bigger, then that matters more than an artificial benchmark-type set-up.

Paul Allen

[doublestar]

Paul Allen ( Microsoft cofounder ) is the key investor in CNET Networks ( pumped $5 million into the company in 1994 to keep it afloat )

Trolled by ZNet

[birder]

This is a typical "imflame the technophiles" hit/ad creator for ZDNet. They do this time and again and ./ always falls for it. This guys conclusion is based on a small (like 1) test case that he's extrapolating to be an all encompassing fact.

Collusion != Identical interests

[Anonymous Coward]

Of course there is some collusion between Intel and Microsoft. But the degree is open for debate. And almost nobody would argue "everything that is good/bad for Intel is good/bad for Microsoft, and vice versa". Their interests are certainly not in perfect harmony.

Keep in mind that the original poster was suggesting that Intel have an interest in making OpenOffice look bad. While Microsoft obviously has a strong interest in making OpenOffice look bad, I fail to see any reason Intel would have a corresponding interest. Simply pointing out that Intel is a part of the Wintel duopoly is a very weak argument.

So long as we're on the topic of slow starts...

[QuietLagoon]

The Thunderbird email client also takes a loooooong time to start up.

OOo Is Not Perfect But Only v2.0

[Anonymous Coward]

Yeah, I recently cut-n-pasted a howto with lots of graphics from my http client in to OOo. Man, was it chocking just to scroll up/down the page. But you know what? It didn't bother me that much because I figure: (OOo Speed)/co$t goes to infinity in speed per buck while (MS Office speed)/co$t has a lower value.

And time doesn't stand still either. OOo will improve with time.

Speed difference

[Anonymous Coward]

"Most of that massive speed difference is due to XML being very processor intensive, but Microsoft still handles its own XML files about 7 times faster than OpenOffice.org handles OpenDocument ODS format and uses far less memory than OpenOffice.org."

I don't know if we should blame XML for the bloat.

MS Office has years of optimization in the look & feel -- and speed is part of the "feel" experience.

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[filesiteguy]

I run KDE (arguably considered bloatware itself) on my laptop with 1GB memory. I loaded an Excel spreadsheed (23K) by double clicking. It opened in four seconds for editing. Not quite sure how to read it (i'm a total newbie) but the process table shows calc as having 135,468 VM Size and 67,700 VM Res.

Opening the same spreadsheet in Windows 2000 with 512MB memory running Office Pro 2003, it took eight seconds to load. According to task manager, excel shows 28,992K. Now I know Excel relies on the MFC so I opened a process viewer. According to that, Excel has 48 DLL's loaded. Um, that seems like a HUGE bit of memory.

Re:Don't let Ou's argument dissuade you from freed

[vandit2k6]

How many people (well average users) actually contribute to the OO even though it's free and open source? Yea sure its open source, so what. I dont care if MS Office is not open source it suits my needs. Oh and besides Office comes free from my school :-).

Re:shoe on other foot this time? weird.

[mrchaotica]

Is it bad that I'm 21 years old and think the same thing?

Now hear this ....

[o_miljac]

"Most of that massive speed difference is due to XML being very processor intensive, but Microsoft still handles its own XML files about 7 times faster than OpenOffice.org handles OpenDocument ODS format and uses far less memory than OpenOffice.org."
Smell a troll here? This aims toward the M$ own XML format being "more efficient" than open XML based format? And using less memory? Bet it is a 10-byte document when opening 7x faster & using much less RAM :-)

Better yet, OOo can recover bad MSO docs

[zooblethorpe]

I've had to deal with some convoluted MSO doc issues, what with complex corporate templates then being mangled by the occasional bug in doc processing tools. Sometimes simply opening a borked MSO doc would cause the program to crash (mostly seen in Word and Excel). This was the only reason I kept OOo on my XP machine at work, as OOo simply trashed all the careful layout positioning of the original (much better in newer versions) -- but when a doc went belly up, OOo was the only thing we had that could open the doc, and then fix it. Simply opening in OOo and saving would fix whatever the issue was, and we could re-open in MSO. Sure, we had to copy and paste content back into an older version if we were worried about borked layout positioning, but we could get the content, which was the most important thing.

For all the other issues, it looks like the OOo dev team understands MSO's file formats better than MS does. :)

www.abisuite.org

[Medievalist]

You're right. Mmmmmmmmmm, AbiWord!

Unfortunately no Abi replacement for Excel yet.

On the up side, there's also no analogue of the mind-destroying PowerPlod.

Can I get an Amen!

[jmorris42]

> The best office package I've used is Lotus Smart Suite. I'd be glad to pay a three digit
> sum for a cross-plattform version (Linux/OS X/Win) of that office package.

I'd drop $99 for SmartSuite without thinking. At $199 I'd think a bit, maybe bitch and moan a little and then pay up. But IBM can't bring themselves to invest any money into it and they can't just set it free either. So it languishes in limbo becoming more obsolete year by year until it's value will slide to zero.

Don't compare apples to oranges

[kylef]

Anyone running OpenOffice on a Mac want to add another data point where MS doesn't have code "hidden" in the OS?

Hidden code, you say? Before you go off accusing Microsoft of a Consent Degree violation, perhaps you should be a bit more careful about what exactly you're comparing. It is extremely important when you try to compare "memory usage" on different Operating Systems that you are actually comparing apples to apples. And since you didn't cite the source for your "7.10" and "9.81" numbers above, I doubt you really understand what you're measuring.

If you're using Task Manager, for example, you will by default only see "Mem Usage" which reports the physical memory (i.e., the "working set") consumed by the process. Even though this metric includes both private and shared pages (i.e., shared code and data segments of DLLs are charged to each process here), it does NOT include pages which still reside on disk (either in the executable images, memory-mapped files, or the system pagefile.

Another common memory statistic from Task Manager is "VM Size" (you have to add it to your column view by "View->Select Columns"). "VM Size" tallies private virtual bytes consumed by the process. Private means that this quantity does NOT include shared/shareable pages like DLLs and memory-mapped files. "VM Size" is sometimes smaller than the "Mem Usage" precisely because shared pages aren't counted. This causes a large amount of consternation to those who don't understand what is being reported, because they expect physical memory usage to be smaller. "VM Size" is the equivalent of the process's page file allocation, since shared pages by their nature are already backed up on disk elsewhere.

Another common memory usage metric in Windows can be obtained from Perfmon (perfmon.msc, the Performance MMC snap-in). From this tool, you can view "Virtual Bytes" of each process, which is the amount of reserved virtual memory for the entire process, including shared pages. It is equivalent to "VM Size" from task manager PLUS shared virtual memory.

So, as you can see, it is not altogether obvious what is being reported unless you really understand the details of memory management on the underlying OS. Before comapring application memory usage across platforms, you need to be sure you're using comparable metrics!

Re:Don't compare apples to oranges

[shotfeel]

Before you go off accusing Microsoft of a Consent Degree violation, perhaps you should be a bit more careful about what exactly you're comparing.

I'm doing nothing of the sort. What I'm saying is that Office app may rightly use the APIs that are part of the OS, and that this code should be counted towards its memory footprint. OpenOffice, wanting to be an open source, operating system agnostic application will avoid using those same APIs. As a result, they have to write and load their own code to do things the OS may normally do.

And since you didn't cite the source for your "7.10" and "9.81" numbers above, I doubt you really understand what you're measuring.

You're right, I didn't. But I did immediately (within a minute) follow with a post that the numbers for all the Windows software (MS and Open office) were from the article.

So basically, my point is the same as yours. What exactly is being measured by these RAM metrics? Are they any real kind of metric on bloat?

Re:Don't compare apples to oranges

[monkeydo]

If MS Office uses shared API's, and OO doesn't, why should that count against MS Office? A tradeoff that OO makes to ensure portability should go in the - column. Portability should go in + column, and may or may not make up for it, but you shouldn't just call it a wash.

Re:Don't compare apples to oranges

[shotfeel]

I really don't know that there is a "correct" way of adding up RAM use. That's my biggest gripe with the article. Is RAM use reported by a utlity any kind of measure for bloat, or how good/bad the software really is?

Re:Don't compare apples to oranges

[Anonymous Coward]

Yet OpenOffice, at least 1.x is slower than MS Office 2000 running under Crossover Office. If OO isn't so bloated then why does MS Office, even with the overhead of an emulator/API translator (pick one) seem faster than OO?

Re:Don't compare apples to oranges

[Ptur]

Using the WinAPI means dynamically linking something like NTDLL.DLL... In the end every program does this because that's how one interfaces with the OS. And is resides in the App's memory space. Or do you know another way to run code on Windows? No. So what's the difference between OO and MSO running on Windows? None. The only difference is that portable apps use some kind of windowing library. And that one than calls WinAPI. Now, can you prove that MS actually included Office code in the OS for the sake of a smaller footprint? Because that is the only thing you can claim when you talk about using a part of the OS that OO doesn't. Quite unlikely, no? Peter

Re:Don't compare apples to oranges

[Jon_E]

It is extremely important when you try to compare "memory usage" on different Operating Systems that you are actually comparing apples to apples.

I'm confused .. I thought we were comparing apples to wintel (or winamd) boxes .. orange hasn't really done a PC in over a decade and went out of business back in May ..

Re:Don't compare apples to oranges

[cerberusss]

Excellent post!! Moderation well deserved. I've posted this on my wiki for future reference (properly attributed of course).

Unfair comparison

[Anonymous Coward]

The article compares the memory usage of "soffice.bin + soffice.exe" (37 Mb) against "winword.exe" (9 Mb).

I think he forgot to add Windows to that account. MS Office requires Windows, doesn't it?

A better comparison would be:

• Windows+MSOffice => hundreds of Mb
  
• Linux+(Gnome|KDE)+OpenOffice => hundreds of Mb
  

Now we can start talking. :)

Bloat? How about this for bloat?

[john83]

For a little comparison, I opened an *.rtf file from my documents at random. It's 14kB as rtf. It opened pretty much instantly with OO and Word both (I've used both since logging on). Saved as *.odt and *.doc, my file is now 30kB with Word. And OO, you ask? Still 14kB.

Research is the application of your own bias to something.

MA?

[colonslash]

Does anyone know if Massachusetts is planning on developing for Open Office? I'm assuming they plan on using OO for their Open Documents, can people corroborate that?

It could make sense for them to - cutting some bloat could make their state employees more efficient overall.

AbiWord v2.4.1 vs Word 2003

[Tumbleweed]

I tried this recently, just to check on initial memory usage, and fired up Word 2003 and AbiWord v2.4.1, typed in a very small line of text into both (same text), and checked memory usage under Windows - and found AbiWord using slightly less than half the memory that Word 2003 was taking. I've not tried having AbiWord laod up any Word documents or anything like that, yet, but it's certainly worth checking into.

Re:AbiWord v2.4.1 vs Word 2003

[Anonymous Coward]

See http://www.geocities.com/typopl/bug5291.html#2005O ct3 for a speed and memory comparison. OOWriter is almost as fast as MS Word. But both AbiWord and OOWriter are getting slower with each new version.

Re:AbiWord v2.4.1 vs Word 2003

[Tumbleweed]

That's an interesting and informative page.

I just installed AbiWord v2.4.1 on my work machine and compared it to Word 97, and it's using more memory (15MB vs 13MB), so I dunno. It seems plenty fast and featureful, so I guess that's good enough to compete with a product that isn't open source or free. *shrug*

I'll have to try OpenOffice sometime, though.

I _really_ wish Evolution for Windows was ready. But since I'm moving to a Mac next year (when they go Intel), long-term, I'm more interested in native OS X versions of software these days.

USE="-java" emerge openoffice2

[phsdv]

I compiled openoffice2 without java and it still huge. I just started openoffice2 without opening a document it uses 119MB Virtual and 39MB res memory.

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5686 paul 15 0 119m 35m 24m S 0.0 4.7 0:04.19 soffice.bin

Loading a 3.6MB csv into Calc changes this into:

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5686 paul 15 0 178m 78m 41m S 43.4 10.4 0:38.89 soffice.bin

A woping 179MB! So no, it is not Java.

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[Anonymous Coward]

I just timed mine -- Writer took 45 seconds to start up on my gentoo linux box running KDE. It's a 1.2 gig with about 512MB ram. Only other things running are Apache and MySQL. It's about 40 seconds until the splash screen even comes up. The memory stats: VmSize = 155,576, VmRss = 62,760. For some reason, soffice.bin is listed 6 times. I don't know why.

To say the least though, it is frustrating to have to wait nearly a minute to open OO. At least if I open it up soon after I closed it, it opens much faster.

Simple solution

[Anonymous Coward]

If OOo XML format is quickly becoming the standard format in the open source world.

So, all we need is a small (preferrably command-line) MS Office -> OpenOffice converter which can be called from Abiword and other small alternatives.

Now, who is the pig??? :D

Statically Linked Libraries

[Anonymous Coward]

To avoid the user having to install umpteen different libs to get a app to run, a lot of pre-compiled applications on Linux ship with statically linked libs. The libraries are compiled into the code. So there is no ".so" Hell for linux users, but conversely the app is a memory pig. I know FireFox does this as well. StarOffice probably uses it's own instance of Freetype, GTK, etc.

If you DLd and compiled the source yourself ( Say, the crazy Gentoo guys ), I suspect the outcomes may be different.

Part of the problem is libraries mature so fast under linux, and even minor versions can have incompat differences. This is unlike windows, where API changes are static for a lot longer period of time. It's been continually biting us on the butt guys, and even if a distro has a 'standardized' set of libs for one release, well, each distro has a different idea of the 'standard' library version. So again, downloads still need to be statically linked to work on all distros, unless distro specific versions are offered.

Wrong tool

[Jerry]

Why would someone use a spreadsheet that allows only 16K rows per sheet to load about 250,000 rows, 13 columns per row, of flat ascii data??? (Actually, there are numbers in the data but they are treated as strings in the spreadsheet to further slow things down.) Clearly, a database is called for. But, even if you were bound and determined to use a pickup truck to haul your 600 acres of corn to market, you should at least use a one that allows for a longer bed. George should have put those 250,000+ rows into only one OOo odt sheet.

An aside, why would one save the spreadsheet as an sxc, forcing users to convert to the odt format in order to test OOo, then have sheet settings which resize the row heights and recalculate the spreadsheet before it opens to the user? Because he has to avoid the one sheet solution.

speed

[philippecmartin]

how long does it take for msoffice to open an openoffice document ?

With the same files?

[phorm]

Most of that massive speed difference is due to XML being very processor intensive, but Microsoft still handles its own XML files about 7 times faster than OpenOffice.org handles OpenDocument ODS format and uses far less memory than OpenOffice.org

From my experience OO can open Office-formatted files. What is the comparison between OO loading an MS-Office format file and MS-Office loading its own format?

Just a thought.

[Anonymous Coward]

For a PhD dissertation, wouldn't you want to use something like LaTeX?

Re:Bloat? Don't talk to me about bloat...

[Anonymous Coward]

Even if it does use 50 MB, that joke (8 MB and continually swapping) is still valid... Consider the overall theme of those jokes -- is it unreasonable to think that person might only have 8 MB of ram free when [s]he starts emacs? Surely it would then be continually swapping...

Re:Problems with OpenOffice. Solution? Fork It.

[Anonymous Coward]

I think the solution to the problem is the forking of the OpenOffice project. Personally, I use only the word processing and the spreadsheet programs. I think that most people only use these two parts of the program...so why not focus on these two? Make them the best. Make each program functional apart from the other. Make each of them lean and mean.

Are you INSANE? It's easy to SAY that, isn't it? Have you _seen_ the code for OpenOffice. My god! *YOU* fork it.

No, better to start from scratch. Maybe borrow some of the code here and there. And yes, that is a huge undertaking -- which is why it hasn't been done. But it would be easier than trying to base anything on OpenOffice..

'system' libraries

[natmakarvitch]

afaik MS developers of various softwares can somewhat better factorize their (common) code because they work more closely, on this particular side, than the 'free world'

the core MS-Windows system has high-level libraries used by all Microsoft applications, therefore (in theory) any software developer working in a Microsoft team is not poised to develop whatever was already coined by another. they cooperate

many functions used by many applications (for example MS-Office) are therefore more or less ready-to-go, maybe even already loaded in the memory occupied by the operating system, when you invoke them. this speeds things up

under a free Unix the people developing system and applications rarely talk to each other. the system provides much less high-level functions (above the libc, X) than any modern MS-Windows and the only teams stuffing more factorized high-level code inside it are working on toolkits (KDE, GTK...).
note: one may argue that those libs are user-space things, please note that I use 'system' in the broad sense
this is, at first glance, less efficient than MS-Windows because there is no central authority pumping functions into a single library used by all apps, therefore many applications don't use them. any given desktop will consequently probably use many libraries providing the same services, therefore many free applications come with their own code to do things already done by other ones.

the MS-Windows approach is theoretically wonderful, mainly because less memory is occupied by various codes doing the same thing.

but it leads to various pitfalls. here are some not neglectable ones:

• a truly useful library is generic and to reach this stage its developer has to code a simple thing (thus reducing the factorization), to bloat it it or to be a genious. in practice they are more and more bloated therefore not bug-free
• any bug in it may cause a failure in any software using it. if nearly all software use it, well...

the bloated high-level services provided by MS-Windows are less stable when the machine simultaneously runs many softwares, thanks to awful side effects spawned by heavy multi-layered codes. therefore most users try to run, on a given box, as few services as possible... loosing a good part of the factorization benefits

from my experience the Unix approach is more and more efficient as time passes. it now provides an environment more flexible, easy to maintain and extend, which extracts more useful power from the hardware

it even gains, after a while, the benefit of the 'central library' approach because efficient and stable libraries tend to gain audience among developers, providing a common ground a posteriori (created for a given client code, then evaluated and adopted for others). on this particular matter the efficient way to do thing wins again: don't try to predefine the whole solution, keep it simple, progress slowly, prefer field-proven solutions

this somewhat reminds me of the 'forked childs' classic Unix trick, for a piece of software, to honor requests for each request: run an instance for each request. this means that a bug will probably scrap a request but not block/stall the whole service, which is a very simple and efficient way to achieve crude but often sufficient software-fault tolerance, albeit it is was ressource-hungry. I write was because it is much more less a system hog now, thanks to some low-level enhancements (started w vfork, copy-on-write and such) conceived after this approach

a Unix system developer tend to adopt simple and proven solutions and then fix their issues. a developer of the MS-Windows system overengineers then tries to the make the whole thing run

from my point of view the choice is a no brainer

WebDSign: thrust the Web by trust

can't belive it

[smutas]

I'm using OOo 2.0 and i can load any document within 10 sec. (centrino 1.6 XP SP2) I'v tried also 1MB word doc (80 pages), with OpenOffice and I can open doc format in 4 sec, when I save it as .odt, I can open it also in 4 sec, as xml OpenOffice need 3 min. OK , I don't care. At first, I think they will improve processing xml documents, but with crossplatform OpenOffice I can work with they own format or directly with MS formats.

Thank's for the Phil 101 lesson

[MooseTick]

"What about the source? Ad hominem attacks are a logical fallacy."

I see you're taking philosophy 101 this semester. Be sure to be on the lookout for other strawman, red herring, and ad hominem arguments. No go back to class.

This reminds me of high-end audio enthusiasts

[Alcemenes]

I can not help but remember when I used to mingle with the high-end audio crowd when I read articles and comments such as these. When CD players became affordable in the early 90s nearly every manufacturer hyped the oversampling rate at which their player operated. This supposedly resulted in fewer errors and better sound although the differences were often unnoticable. The technical specifications quickly became a guide as to the overall quality of a piece of equipment, whether it was a receiver or a CD player. For most people this was sufficient and they quickly snapped up commodity bargins through various catalog and retail outlets. High-end enthusiasts approached selecting stereo equipment somewhat differently. An audiophile would often visit a local stereo boutique with a couple of their favorite recordings and spend several hours listening to different pieces of equipment. The technical specs were important but not as important as the overall sound of the equipment. Of course there are arguments regarding whether or not there is any merit to this but that is not the point I am trying to make. I think the way people approach their choice of software is very similar to how they would approach purchasing a stereo. Many folks choose the Microsoft option because it's a name they trust and they can easily be impressed with a whole bevy of statistics they may or may not understand. Folks who opt for open source products such as Open Office tend to look a bit deeper. The number of overall features matter but not so much as whether or not they can use the same software on multiple computers, on different platforms and architectures and whether or not they can easily exchange documents with their colleagues. That last item can be a bit of a sticking point but overall I think the analogy fits. Is Open Office bloated? I guess it depends on how to define bloat. Feature-wise it feels more sparse than Microsoft Office, resource-wise it tips the scales a bit higher and feels more sluggish. I still use it though. I can use it on any computer I own, it allows me to export to PDF much easier and the price doesn't strain my budget. My parents on the other hand are still very new to using computers and the consistency of the interface offered by Microsoft Office suits them better and my mom can purchase a book about Microsoft Office much easier than for Open Office although that is changing. I think the point I'm trying to make is that I think people are going to continue to use Microsoft Office because the "numbers" look better. As Open Office continues to mature I think this will change, especially as people become more budget conscious.

But the Important Thing is...

[baronvonwalz]

That Open Office is free, where as Microsoft Office costs $300.

Office is older...

[Anonymous Coward]

I think the most simple explanation is that MS Office is much older than OOo. But this makes me think OOo is even better if they can catch up to MS so fast...

Re:shoe on other foot this time? weird.

[ArmorFiend]

So is it better to spend $X00 on M$-office, or $X00 on RAM to make OO perform better? Hm....

Gnome Office

[linforcer]

I've been using Abiword and Gnumeric for some time, and I like it much better, especially on my Gentoo system where everything is compiled from source. I guess these pieces of software aren't as featureful yet (they don't have easy integration, and still lack the rest of a productivity suite, but at least when I start them, they, start.

(Of course by that logic, I could get some Free DOS-thing and run WordPerfect and Lotus 123)

Because loading happens so often...yeah, right

[neonfrog]

A real test would involve 2 different clean installs of Windows: one with Office (and nothing else), and one with OO.o (and no Office having ever been installed) and maybe JVM. Respectively, have Word or Writer auto-launch on boot. Begin typing. Now show me the time from power up to printing a standard business letter that you compose yourself. And total system memory usage (not application only) as a percentage of total available.

What, boot time isn't fair to include? Eliminates the "pre-load" camp's arguments quite handily. Even better, have the system auto-load the specific file for the "file-viewer" benchmark this blogger is doing.

How often do you load a word processor every day? If you are an infrequent user, the overhead is ultimately small because usage is low. If you are a power-user of the office suite than the impact is greatly reduced because it only gets loaded once. Then it gets USED. Where are the usage benchmarks?

Show the numbers for load time as a percentage of your day. Value that time in cash. Weigh that cash against the cost of the two systems. That is assuming you have no OSS or anti-Microsoft agenda = Priceless. THAT's a better benchmark.

This blogger learned about ProcessExplorer. Yippee. He should also learn that a real Office benchmark includes actual usage and creation of content. His benchmark shows the efficacy of the two programs as file viewers only.

What Bull

[NetRAVEN5000]

Somehow I think there's something wrong here. Maybe part of Office is built into Windows or something, because I'm using Linux, and it takes 2 seconds before the OO.o 2.0 startup screen comes up and then another second before the actual office suite itself comes up.

Or could it be that MS has something built into Windows to hinder Java? I know OO.o uses Java, and that wouldn't exactly be the first time MS used Windows to hurt competitors. . .

Another problem...

[Dster76]

...I just started using openoffice for the first time. The word 'internet' gives me a red squiggly. That pisses me off more than bloat.

Re:Another problem...

[Anonymous Coward]

Yeah, it's supposed to be 'Internet':)

Re:Another problem...

[Anonymous Coward]

Poor spelling skills would piss me off too.

Re:Another problem...

[Anonymous Coward]

That *could* be because "Internet" is a proper noun, and requires a capital letter.... just a thought, of course.

Re:Another problem...

[Dster76]

wow. I learned something. thanks anonymous cowards! "The Internet is the largest internet in the world" - Source: The Free On-line Dictionary of Computing

Compression

[Dwonis]

Does MS compress its XML? OpenOffice's format is XML *inside* a .zip container.

bxml vs. XML/zip

[shis-ka-bob]

There is a binary XML format called bxml. If OpenOffice were to use BXML for the internal representation of the data, there is reason to believe that parsing XML documents could be significantly faster. In fact, there are performance comparisons using XML from OpenOffice.org's Writer. It looks like a speed up of 4x to 6x would be possible if bxml were to replace xml. This is still an open format and it can be (trivially) converted back to XML. This would be a natural way to have the portablity of XML with an efficient binary data strructure. And it can be done with open source libraries, hopefully with only small changes to the existing code.

Re:bxml vs. XML/zip

[Doc Ruby]

Some binary representation of XML might be better than XML. But the point is that no single data format, or even metaformat, is going to be optimal for every arbitrary app's internal processing. Internal data structures are best designed to suit the internal algorithms and other app-specific constraints. XML, binary or otherwise, or any other generic format, will not have that "custom" suitability. So all these formats have their place. But there is no single format that fills all the places. Data interchange between apps has enough benefit from interop that losses of specific processability don't ruin the XML "value proposition". Partly because different apps on either side of the interchange mean that one's optimized format is inappropriate to the other; XML at best can split the difference. But once inside, the native formats are best used.

Re:bxml vs. XML/zip

[armentage]

I believe Office apps use some sort of a memory mapped file approach. They don't load the files into memory so much as they use the file on disk as a sort of database. This is why Office apps don't grow so much as you make your documents bigger (especially Excel). The OS handles efficient paging, and the app only holds on to fast hash index of where to find information on the disk (which might come out of the disk cache or be read from the disk on each access).

You can't do this with XML, and this is why XML sucks (for computers) to "work" with. It might be a good way to move data around, but saving document data in XML and trying to use it at runtime is always going to be a bad idea. With XML, you need to load this giant ASCII file and convert it into a binary representation before you can do anything with it. Unless you write your own VERY smart XML parser, most current tools will want all the data in memory and want to store all of the result in memory. POW, you've doubled your apps memory requirements at load time, and you're going to have to hold on to a copy of the converted data at run time (or drop it to disk as a tmp file).

Pathological test case... why has nobody noticed?

[CCW]

If you look at the test files, they are HIGHLY redundant, consequently they get zipped extremely effectively .ods is about 3MB but uncompressed xml version is around 192MB.

Based on the memory footprint, Excel is clearly not loading the entire file into ram, or windows is not counting it toward excel. OO on the other hand, not only has to uncompress the file into memory, but hold the entire 192MB file in memory since it is in a compressed form on disk. That comprises the bulk of the difference in memory footprint. Nobody seems to be counting the uncompress time done on the excel file before it is opened either.

This is clearly a pathological test case - however it does lead to a couple interesting strategies: OO could optionally save files uncompressed if the user thinks opening speed is more important the disk space. OO does use more memory and is slower than the Excel app, and this should be addressed for 2.1

The correct response to somebody like George Ou is to say "Thanks for finding this nice test case" and then figuring out whether it really matters, and if so how to fix it.

More meaningful measurements

[slayer99]

I produce and proof read quite a lot of technical documentation. The length of time is takes an application to start is nearly meaningless. I'll do that once in an 8-12 hour stint of work.

Metrics along the line of "repaginate a 2000 page document after increasing the generic font size by 1 point including custom headers and footers" would give a better idea of performance loss or gain against Micrsoft's offerings.

but it runs on my preferred OS

[dindi]

Besides the fact that it runs on my preferred OS as opposed to MS office, i am pretty satisfied with the speed of all the versions I used. It id bloat for me for sure, but opens whatever I throw at it and usually it runs for 1hours tops a week so I do not care really.

I have to admit that i have to deal with very simple tables which I tend to dump into csv, or a mysql table so I can manipulate it without clicking like an idiot .... also I convert everything into rtf or TEXT and avoid any other format as much as possible..

I do not use the presentation maker and the draw at all so I would actually not mind having a thin version (maybe there is one, but I just apt-get openoffice usually at night and I have it in the morning :) )

I run an old machine for my desktop stuff: amd xp2000 with 768m and 2 old nvidia fx-dunnowhat, so my "benchmarks" are not relevant anyway nowadays :)

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[clear_thought_05]

You are using 2 DIFFERENT computers. What can you conclude? Different CPU's, different memory speeds, different harddrives, etc. You are comparing apples and oranges. ...

Oh yeah, well I opened a 100K text file on my 3Ghz Linux box in 0.5 sec, but on my 8086 with 256K of memory it took about 45 seconds to load off of the 5.25 floppy disk. See my point?

How does 48 DLL's conclude you to thinking that's a lot of memory? Wouldn't that be a lot of libraries? Run 'ldd' on some of your favorite Linux apps for some side to side comparison.

And just to be helpful in people creating useless statistics, I opened a 540Kb excel sheet in about 4seconds on excel2003, Win2K, 512MB, 1.6Ghz laptop. The time it takes to load from a 10Mbit LAN: 6 seconds. I'll take it home to see how long it takes to load on my desktop with OOo

article forgot to calculate the cost of FREE

[Anonymous Coward]

FREE vs few hundred dollars...

PDF size is definitely bloated, ODF is not

[MLamar]

As a quick-n-dirty test I used a PC running Windows 2000, OpenOffice.org v2.0 and MS Office 97 to build .pdf files of quotes submitted to me as an Excel spreadsheet (one printed page) and a Word document (8 pages). The .pdf engine used with word and Excel was Software995's pdf printer driver. I accepted the default settings for pdf creation in both OOo and MS Office apps.

The OOo spreadsheet.ods was 16K, the Excel spreadsheet.xls was 24K in size. The OOo document.odt was 38K, the Word document.doc was 131K in size. In fairness, the documents did contain VB macros that the .odt file did not save. The OOo spreadsheet.pdf was 120 KB , the Excel spreadsheet.pdf was 32KB in size. The OOo document.pdf was 200KB, and the Word document.pdf was 100KB in size. I will check PDF creation on these same documents from my Linux box at home to complete the comparison. My curiosity will then be satisfied.

To me, the expense of archiving data files is of greater interest that the expense of RAM needed to support the software. I must archive documents for many years. Disk space requirements are constantly growing, while the RAM in each PC is reusable.

Of course OpenOffice is bloated

[defile]

didn't you see the word "Java" in there?

Oh, but computers are getting faster all the time! Right?

Provide OOo to Linux-Win w/NoMachine

[j0ebaker]

Here's an Idea to speed up OpenOffice deployment.
OK, buy one big server for your network. Install Linux.
Set your use flags.
Install NoMachine http://www.nomachine.com/

Install NoMachine clients on the Windows and Linux workstations.

Create icons on the desktops that will launch the application from the "Big Server" to the clients.
Walah! You have a method for improving the deployment speed of new applications throughout your network from a single place.

Let's look at Gentoo's approach of compiling OpenOffice to see if there are ways of increasing the amount of shared RAM that is used by OpenOffice. Maybe making it less of a static application would help. Then another way to improve execution speed it to introduce a tool called prelink.

Hey, at least we have the application. It is solid and it is free. Now you know ways to deploy it without upgrading all your workstations.

-Joe Baker

Maybe George Didn't Use Enough Test Files?

[dlapine]

I'm sure that George did a full range of tests, of files from size ~10K up to the 100MB monster. I'm not sure why he didn't publish the results of the other files, but surely MS Office showed the same disproportional speed increase and lower memory use for all types of spreadsheets and word documents...?

I read the blog entry, and this looks like an attempt to report the worst case (corner case) benchmark, for whatever reasons George might find useful. A fuller set of tests and results might make his case that MSOffice is better a little more convincing.

Hard to be convincing about the usefulness of MSOffice over OpenOffice, if you're going to ignore the fact that MSOffice isn't available for some portion of the users, and choose to report a single data point as if it were conclusive.

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[filesiteguy]

Oooh, sounds like somebody is a bit testy. I'm just posting what I see. I never posted conclusions. Take my numbers for what they are.

http://johnhaller.com/jh/useful_stuff/portable_ope

[Anonymous Coward]

http://johnhaller.com/jh/useful_stuff/portable_ope noffice/

You got to love free software, if you don't like it the way it is now. Addapt it to your own needs, or have someone else to addapt it

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[clear_thought_05]

Your post really is meaningless. Do the test on the same machine and then post, otherwise you are polluting this already hopeless comparison with more bad data. And I still have no idea what your DLL analysis meant.

Re:Bloat? Don't talk to me about bloat...

[Billly Gates]

Bloat vs features.

Openoffice is behind MS Office and is very bloated many times over. I believe its design if you ask anyone who profiled it and knows about operating system internals such as threading and object loading/unloading.

Re: My Results Differ...

[jcole]

-My specs-
jcole@jcdebian-p4:~$ uname -a
Linux jcdebian-p4 2.6.12-1-686 #1 Tue Sep 27 12:52:50 JST 2005 i686 GNU/Linux

jcole@jcdebian-p4:~$ grep name /proc/cpuinfo
model name : Intel(R) Pentium(R) 4 CPU 2.66GHz

-Time-
jcole@jcdebian-p4:~$ time ooimpress
real 0m6.170s

jcole@jcdebian-p4:~$ time oocalc
real 0m5.537s

jcole@jcdebian-p4:~$ time oomath
real 0m4.928s

jcole@jcdebian-p4:~$ time ooweb
real 0m6.321s

jcole@jcdebian-p4:~$ time oodraw
real 0m6.208s

jcole@jcdebian-p4:~$ time oohtml
real 0m3.792s

jcole@jcdebian-p4:~$ time oowriter
real 0m5.741s

jcole@jcdebian-p4:~$ time ooffice test.sxc
real 0m6.170s

jcole@jcdebian-p4:~$ ll test.sxc
-rw-r--r-- 1 jcole jcole 13196 2005-10-27 11:44 test.sxc

jcole@jcdebian-p4:~$ time abiword
real 0m2.745s

jcole@jcdebian-p4:~$ time gnumeric
real 0m2.343s

jcole@jcdebian-p4:~$ time /opt/cxoffice/bin/wine --check --bottle dotwine --cx-app "C:////Program Files////Microsoft Office////Office10////EXCEL.EXE"
real 0m3.514s

-Memory-
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
25721 jcole 16 0 138m 49m 33m S 0.0 4.9 0:03.89 soffice.bin

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
25888 jcole 16 0 1075m 12m 7976 S 0.0 1.2 0:00.89 wine-preloader

Also Firefox

[jridley]

Not in terms of bloat, but in terms of slow processing. Firefox's Javascript is HORRIBLY slow in comparison to IE's. I have an intranet site that allows viewing of somewhat large tables of data (up to a few thousand rows), and there's javascript code embedded that allows filtering & sorting of that data. Change a selection in IE and it's done in about 2-3 seconds; in Firefox, the same action can take 45 seconds or more.

I still prefer Firefox, but I use IE if I need to view a large dataset on that page.

Just goes to show, even after functionality is there, there is still a lot of room for improvement in many OSS apps.

Another comparison: AbiWord vs Word vs Writer 2.0r

[Burz]

Someone should mod the parent way up.

See http://www.geocities.com/typopl/bug5291.html#2005O ct3 for a speed and memory comparison. OOWriter is almost as fast as MS Word. But both AbiWord and OOWriter are getting slower with each new version.

Please, re-read

[QuaintRealist]

I don't usually respond to this sort of thing, but I think you genuinely misunderstood - You quoted my comment, but did you read it? I do not claim any evil/secret/unfair advantage for Microsoft, but simply wish to point out that they have experience in coding within their own paradigm, which works to their advantage.

Kudos to the folks at OpenOffice.org for the work that they have done - it's a steep hill to climb, and they have made a lot of progress, relatively quickly.

And folks, let's not criticize their coding skills unless you are a coder yourself and can do better yourself...If this describes you, then hey, forget the critique and just pitch in!

Re:Please, re-read

[armentage]

Well, I re-read your post, and I still take issue with the way you phrased that point. That it's "their" platform has nothing to do with why Office runs so well. It's all the same hardware after all. You can only eek so much out of it, and with apps like Excel where its all about memory utilization and clever coding, there isn't much about the OS that really matters.

Someone else mentioned that Office runs great on OSX - we know that OSX is essentially your typical UNIX variant with a fancy GUI system. It's not their OS, but MS still beats OO.o because they code better.

Re:Please, re-read

[Anonymous Coward]

I don't usually respond to this sort of thing

Oh, that's nice. "This sort of thing"? The parent poster crafted a reasonable reply to what was an unreasonable post, and somehow you twist this to make it sound like he's the troll? Pathetic.

You quoted my comment, but did you read it? I do not claim any evil/secret/unfair advantage for Microsoft, but simply wish to point out that they have experience in coding within their own paradigm, which works to their advantage.

Don't accuse people of not having read your post just because they take exception to the way you phrase your trollish points. MS has no magical way of making their software run faster. The simple point is that Office is better coded -- it runs much faster than OO.o on Mac as well, for christ's sake. Just because your brain would explode of cognitive dissonance if MS had made a genuinely faster product doesn't mean it isn't so.

Kudos to the folks at OpenOffice.org for the work that they have done - it's a steep hill to climb, and they have made a lot of progress, relatively quickly.

Yeah, who claimed anything to the contrary? Listen, not everything is some sort of personal attack on you or the OSS community. We're just comparing two applications, that's all there is to it.

If this describes you, then hey, forget the critique and just pitch in!

I'm so fucking tired of this response. So what if I happen to be an excellent coder? I want an Office application, not write one. If the response to "why should I use OO.o when MS Office is so much faster?" is "write it yourself" then not too many people will care, will they? When someone wants to edit a document the last thing he needs to be told is to write the goddamn code himself. Maybe they already have a life, work and hobbies? Maybe they don't want to be told how they should spend their spare time? Maybe, just maybe, they just want some pointers on which office software to use?

Christ.

Re:Please, re-read

[stressmagnetchick]

Put the espresso down and settle yourself. He (or she) wasn't accusing you or anyone else of not reading his (or her) post. He (or she) simply asked if someone had. Trollish points, indeed. By now even you really ought to understand the difference between a troll (e.g., gnaa, penis birds, page-wideners, etc.) and flamebait.

Re: My Results Differ...

[Anonymous Coward]

I get similar results here. 6 seconds is not so bad to start OpenOffice. So it's 2 to 3 seconds slower, so what? heh.

Of course MS is faster . . . .

[denverradiosucks]

This is somewhat obvious to me that Excel would run faster, considering it is created by the same company that created the OS it's running on, and it doesn't have to do any VB to java conversions like OO does. I am sure OO would run much faster on JDS and solaris than MS if MS was ported to Solaris. What's the point? I think the 'bloat' numbers are a bit exaggerated (been running OO 2.0 on my XP system here at work for a week, and it runs great). Just now opened a good sized Excel file in OO (charts, formulas, macros, etc) and its using 64 mb right now. With Excel and the same file, 12MB. It's like being a great golfer and creating a golf course and bragging about how you get lower scores than anyone on that course. Big Deal. OO 2.0 is SUCH an improvement over previous versions, I am very pleased with it.

Re:"Essentially" the same data? MOD PARENT DOWN

[cosinezero]

So what you're saying is that bloated software gets un-bloated if you give it more memory to play with?

Why spend more money on hardware if the computer you have is perfectly fine? Don't answer that - the guy above me with the checkbook is already shaking his head NO.

Poor memory management is a sign of poor quality code, or worse, poor planning cycles in development. Better hardware doesn't make your application better.

intel admitting defeat?

[turgid]

Indeed. Intel should be pushing OpenOffice, because nothing makes me aware of how much I need to upgrade my processor like starting OpenOffice.

Are you sure? intel would be admitting that its processors are feeble and that you should buy an Opteron or Athlon 64 from AMD. After all, Sun does all its development on 64-bit AMD and UltraSPARC nowadays (with the emphasis on AMD becuae it's cheap, plentiful, and runs cool).

Listen to an old timer slackers... Live to learn

[mcdtracy]

We had a similar discussion regarding Novell and TCP/IP...
TCP/IP was PROVEN to be slower, used more memory and cost more to support.
Score ONE for the market leader.

We revisted the discussion withg VMS vs Unix.
VMS was PROVEN to be faster, more resource efficient and delived a safer support value system.

When they suggested that we should drop our elegant and highly reliable SNA network
for the Internet, we did the benchmarks and showed that security was impossible
to guarantee for critical data when we didn't control the WAN. Don't GO there...

Product are market leaders for a reason... because they sell more than anything
and these "try this/it's free" proponents will never hire enough seasoned resources to
out-code a market leader that has a revenue stream on the order of Billions/year.
It's a simple investment verity! One-Billion downloads at $0 is $0 in revenue.
It's worse than that... who pays for the servers and downloads. Think ROI!

Why we need to revisit this issue again and again is beyond me.

Some people just never learn from history. The author of the article
seems to have some experience in predicting the best use of
IT investments.

I need to get back to my Cobol account app that needs new patches
for our brilliant Web/3270 frontend release.

You slashdotters just need to learn from your wiser, older IT system analysts
before we die off. It's almost too late! There is no such thing as a free lunch...
Lunch doesn't just grow in trees. Your lunch gets cooked in factories like the
think tanks in Redmond and Armonk and Cupertino. Chao.

Crazy...

[LeonGeeste]

My post shot up to 4 at first (probably 5), then dropped to -1 and now has oscillated between 1 and 0. (Wouldn't it be cool if moderators could get together, cancel out all the conflicting moddings, and then apply their mod points to other posts?)

I'm just not going to say anymore on this, except that the thing about "running laps around MS" was a general evaluation, not necessarily concerning the speed and storage issues. OO.o is so much easier to use for me, I just love it. That's why I made the incriminating "lap" comment.

Yes its bloated

[nurb432]

However, it does do the job and is still free.

That is a fair tradeoff in my book.

Re:Children, grow up and admit that OSS isn't perf

[Dcnjoe60]

While I agree with your premise that "bloated" and "slow" are always bad, I'm not sure that using a 200+ MB file is representative of how most people use Excel and/or OO. Therefore, if someone found some type of atypical data set that would by it's very nature make either Office or OO perform differently than the average user would expiernce, it's hardly appropriate to use either bloated or slow to describe the results.

A much more representative test would be to use files that are more typical and common. I doubt that the majority of users have such large spreadsheets and documents, at least not for their normal usuage (and if they do, they really should look at using a database instead).

I'm not making a case that OO is better than Office or vice versa. However, using the testing methodology as described, my Ford Focus is slow and bloated because it can't go as fast as a Formula 1 race car. And that would be true, if your criteria was just on the top speed. However, that's not how the typical automobile is used.

So, yes, the Formula 1 is faster and leaner than the Ford Focus, but since the speed limit is far below the top speed of the Formula 1, does it really make a difference? Likewise, Microsoft Office can load more quickly, some huge data file that the average person will never have, but does it really make a difference?

Yes, OSS isn't perfect, however it's strengths (and weaknesses) are not measured in load times and/or memory footprints.

Oh really?

[jotaeleemeese]

Me here using OO.org for the best part of 3 years for my Open University assignments and have not noticed.

This is a non issue for anybody that looks at the cost of the alternative, not only in monetary terms, but also in terms to access to your own data.

I don't care if it takes a bit longer fo OO.org to do this or that task, at the end I have a document that I am pretty certain I will be able to open and manipulate in a few years time, and I will not have to pay anybody for such privilege if I don;t want to.

Keep putting your data in the hands of the beast, eventually you are going to be deprived of it and will be charged to access it when you more need it.

MOD PARENT UP

[bmalia]

Tired of all the Java FUD.

SSE2 floating enabled?

[dedded]

I ran a quick, one-datapoint test. Opening a Word document was as fast in OOo2.0 as in Word, but opening a large Excel spreadsheet in (an already running) OOo2.0 was very slow. Most of the time I had a progress bar labeled "Calculating..." at the bottom of the screen, so I don't think XML parsing was the culprit.

On my Fedora system, the OOo calc package lists i386 as the architecture, so I'm wondering if SSE/SSE2 floating point is enabled. The old stack-based floating point is slow.

/Dan

Re:Bloat? Don't talk to me about bloat...

[Anonymous Coward]

EMACS Makes A Computer Slow

Esc-Meta-Alt-Ctrl-Shift

You get what you pay for; simple as that.

[cruncy_eyeballs]

Subject really says it all. All this open source stuff is for the birds, though I'll admit to raiding a few open source projects for ideas. But damned if I'll use there code. Most of it is C or heavily abused and obfuscated C++. If you want something that's fast and works great, no matter what OS; then you're going to have to pay for it. There's a whole different paradigm when you're writing something to earn a living and writing something for free. You just don't have the same accountability... period.

Bah

[PeelBoy]

Is this not exactly what they mean by bloat? I mean if it were the other way around we'd all be bitching about bloated MS software loading all of office up when all I freaking need is excel. It's retarded. Opening up Word, Excel, Etc.. is NOT a fair comparison at all. If all I need to do is 1 thing why should I have them all open? Well with Ope nOffice it sounds like I don't have much of a choice!

Not that this will stop me from using it, but I hope it's something that gets fixed in the future because it really is retarded.

Although.... Correct me if I'm wrong but StarOffice wasn't open source, right? And Open Office is based off StarOffice.. How does this translate into open source software being bloated? I mean it wasn't even fucking open source for most of it's life.

Lets compare Excel to Gnumeric. That's a MUCH better comparison.

Graphs.

[Yobgod Ababua]

My point was more that you're unlikely to increase any understanding of your data by generating said 10,000 point graph, as only about a tenth of them would actually be visible in the output. Similarly, you're probably not going to learn anything useful by looking at a 10,000 row spreadsheet with your eyeballs either. That's too much data to absorb at once, it needs more cooking before it's ready to be displayed in a useful manner.

I'm not making any excuses for OOo. It -is- noticeably slower and it shouldn't be.
I haven't thrown a profiler at it yet to help determine -why- it's slow, but we should.

Re:Children, grow up and admit that OSS isn't perf

[Theovon]

If you read Gödel, Escher, Bach, you'll read about this idea of a record that cannot be played by a certain record player. If you change the record player to play that record, you can still find another record that the new player can't play.

So, yes, I agree that it's certainly possible to make a spreadsheet that works well with Excel but not with Calc. And I also agree that it may not be a fair test.

But I was more interested in the memory usage for empty documents.

Have you actually tried OpenOffice 2.0?

[jgoemat]

2. It is a known problem that OOo takes a while to start. Staroffice (at the point when Sun bought it) was made by a German company. Most of the internal functions are named in german, and use abbreviations that are not obvious. The fact is that each version of OOo has been getting smaller and faster. OOo 2.0 is the same. If you run OOo 1.1.4 and OOo 2.0 side by side on windows, the 2.0 version uses about 10MB less memory when both have nothing open.

Calc and Writer both open up in about 2 seconds for me... Excel opens up a bit faster and Word takes a bit longer (about a second). If I turn on Quickstart, the times are under a second... Maybe the last version of OpenOffice had that problem, but this one doesn't for me (Windows XP, 3.4ghz P4).

Re:Children, grow up and admit that OSS isn't perf

[Theovon]

Oh, and I neglected to comment on your last point ("Yes, OSS isn't perfect, however it's strengths (and weaknesses) are not measured in load times and/or memory footprints."). This is the crux of my argument.

FOSS advocates use load times and memory footprints as a means to measure proprietary software, and they do it all the time. And what I'm saying is that since those things matter then, then they also matter now, and the comparison has to be made fairly for all software.

Re:Children, grow up and admit that OSS isn't perf

[KayosIII]

Except for one thing. The OSS world also has software that is not Slow and bloated. This software does not necessarily play well with Microsoft Office and is not necessarily available for non free OS's at this point. So some of us have to deal with Microsoft Office users - using OSS software we can have either light/fast or MS Compatible at the moment - Not Both. On my box (Mepis Linux, 2.5gig Athlon, 1 Gig ram) (I just switched java off - thanks for the Heads up) OOo 2.0 loads in an acceptable time frame. I haven't got any particularly large documents to test the file loading/saving slowness. But I am happy to report that one of the big speed bugbears from OOo 1.1 series is gone - loading charts into Calc. The generally editing is quite snappy too. Would I like to OOo to be faster and use less memory. Yes I would. For the 2.0 release was it more important than having better MS Office compatibility. No it wasn't (not for me anyways).

Maybe this is why...

[J. Random Luser]

MacOS 10.4.2 OpenOffice.org_2.0.0rc3_051016_MacosxPPC run Calc:
process real-mem. virtual-mem.
droplet 8.5M 36.5M
X11 0.8M 27.1M
Xquartz 21.8M 168.9M
quartz-wm 3.1M 107.6M
xterm 5.5M 26.9M
sh 0.7M 27.2M
tcsh 1.0M 31.1M
soffice.bin 73.4M 282.1M

All of these processes were started by OOo_Calc. I don't have MSOffice on this machine so no valid comparisons are possible. I observe that OOo is a slug to start from cold, but once up and running seems snappy enough opening, saving docs. I do notice on cow-orkers Macs that MSOffice is running away to talk to the printer every time it opens or saves a doc....

its the java dumbass

[sl4shd0rk]

Everyone knows java is a friggin resource PIG. Add a bunch of threads and watch the memory fly! I happen to like java very much, and use it wherever I want cross platform portability. That in itself can be enough of an advantage if that is what you need. Not the perfect solution for everything, but it has very good advantages depending on the situation. I'll gladly add another gig of welfare grade ram to a machine if I need the portability. If your're running windows and openoffice, you have simply shot yourself in the foot twice. Microsoft never wanted java, never wanted to play nice with java, and as far as I'm concerned, still doesn't. Buy some crutches and accept your inevitability.

I Rest My Case

[Anonymous Coward]

MS office ver:200.0
Open office Ver 2.0
  Thanks for coming out

Gunillablue

Size Matters

[nagalman]

I saved this file in .ods and .xls format in OpenOffice 2.0. I opened up the .xls file in Excel (and it does load MUCH faster) and saved it again (again MUCH faster than OO 2.0). File size results: .xls = 52,239,872 (49.8MB) .ods = 4,045,860 (3.85MB)

Sometimes size does matter :)

Re:Bloat? Don't talk to me about bloat...

[Anonymous Coward]

What happens if your keyboard has Left, Right, Hyper, Super keys? Do we need to create a new text editor, Lermscash? Or is Emacs sufficiently adaptable that it can be Lermscash, too?

Back in my day...

[broody]

we couldn't even open word documents and liked it!

Re:Children, grow up and admit that OSS isn't perf

[Anonymous Coward]

Okay: OSS isn't perfect. I said it.

Nothing is perfect. There are problems with everything. As a rule, I prefer the problems you get with free software, rather than the different problems you get with proprietary.

In the case of OOo, the bloat dates back to StarOffice. I actually tried to use StarOffice 5.x and it was sooo painful on my K6-III/450 computer.

OOo these days is better than StarOffice was but still far from perfect.

I really like both AbiWord and Gnumeric, and I use those by preference. I never use OOo Calc for anything; Gnumeric is all I need. But OOo Writer is much better than AbiWord at opening old Word documents and doing the right thing with them.

Of course my current computers are wayyy faster than my old K6-III and that doesn't hurt either. OOo, despite the bloat, meets the "Good Enough" standard.

By the way, it hass to be said that MS compilers make better code than GCC. If we could recompile FOSS in better compilers it would be better (but OOo would still be bloated, separate problem there). But that's why a comparison of GNOME or KDE vs. Windows 98 made Windows 98 look good on memory issues. (These days, XP burns memory and the FOSS desktops look much better in comparison.)

Re: Linux/MacOS/FreeBSD/Solaris Benchmarks Please!

[k12linux]

OOo v2.0.0 uses 15.1Mb (-/+ buffers/cache) RAM when I start it on Linux.

Startup without loading a document was 6.5 seconds first time and 2 seconds once cached. Yeah.. I think I can live with all of that.

Has anyone tried running both under wine?

[DFJA]

It seems that the only way to discount the tricks that Microsoft play to favour their applications would be to run on a non-Microsoft platform. As MS Office is only available for Windows, this would presumably mean wine (or Crossover Office). MS Office works under wine, I don't know about the Windows port of Open Office (the only fair comparison really) but as both it and wine are fully Open Source, it should be possibly to make it work.

Just curious, I've not tried myself.

Re:Has anyone tried running both under wine?

[Hanzie]

There is an MSOffice for OSX, I think.

I agree, OOo is bloated

[noamsml]

That's why I often prefer to use Abiword. I often don't need anything other than a simple word processor wit basic stuff like headers, highlighting and spelling corrections. for math homework I use OOo, but I am checking into using some advanced app like Texmacs.

Excel is *not* excellent

[adoll]

Excel is a serious problem for people like me doing circulating load calculations in process engineering. See my papers here and here. It is OK for chequebooks, but don't expect to design a copper smelter using it (use an ancient ver of 1-2-3 instead).

To be blunt, the guys who wrote the Excel GUI got an "A" in computer science, but the guys who built the calculation engine only got a "C+". To be a truely great spreadsheet, Excel must:

• Use backward chaining to iterate circular calculations
• Not invert singular matrices
• Put in a more robust statistics package, although this may be a sub-set of the matrix math problems.

Any engineer who gives me a calculation done in Excel using circular reference calculations had better be prepared to get his butt roasted. I've had 10Mb files modelling a copper smelter that converged to a wrong answer - that's unacceptable given that the same calculation saved as a 1-2-3 file converged to a correct answer in 10 seconds using Lotus 1-2-3.

-AD

Re:Excel is *not* excellent

[Anonymous Coward]

remember not to put excell to do matlab work...

Re:Excel is *not* excellent

[xenocide2]

Holy smokes batman! Excel is pretty nifty, and my roommate's demonstrated some rather interesting techniques for modelling heat transfer over a copper plate using iterative techniques. But lets be honest, Excel is about a good of a scientific calculator as it is a database.

Like your paper mentions, order of evaluation is critical for correct results. What isn't laid out in the paper is the natural consequence of floating point representations: error. Errors might seem small for any given calculation, but when you're iteratively calculating a solution a thousand or more times, I suppose it's not impossible to wind up with signficant error. The important question I can think of that doesn't appear to be addressed in your paper is: how to determine when error though this method becomes signficiant? Relatedly, can/should a PE use these calculations in certifying anything?

Partly it sounds like you wish matlab was better at visualization ;)

Re:Excel is *not* excellent

[Anonymous Coward]

Excel for scientific engineering? Try Origin (from originlabs), it's best out there.

Re:Excel is *not* excellent

[Lady Jazzica]

Here's a link: OriginLab

Re:Excel is *not* excellent

[adoll]

I've never had a big problem with cumulative error. Typically the floating point errors are in the range of 10^-12. Not significant for the number of iterations I run.

Regarding Professional Engineers using _any_ software ... this is a pet peeve of mine from way back. I don't trust black boxes. I also don't trust spreadsheets until they have undergone a lot of truthing. In order to use a spreadsheet in a formal calc where there are circulating loads (or other circular calculations), I want to see "double entry accounting" and some "sensibility checks" that the results are approximately what was expected. I've seen stats that between 70% and 90% of spreadsheets contain errors, and my experience says that is about right.

Using black-boxes in engineering calcs is a bigger problem that almost nobody in the chemical engineering business takes too seriously. In short, the calculations done in programs like Aspen, Matlab and other simulators will eventually run into any of the following problems: boundary conditions they can't handle, errors in databases, and extrapolations beyond what is reasonable all can give horrific error results. How does the average (junior) engineer using black-box software know that when a model is unsuitable? How often does it tell you that "the model for bitumen doesn't work in the presence of water"? At least if I build the model for bitumen from scratch, then I should know the limitations of it. Then the danger is incorrectly programming the model rather than a hidden model limitation.

-AD

Re:Children, grow up and admit that OSS isn't perf

[lkeagle]

Nice Straw-man...

People who make speed comparisons are usually making them with regards to some aspect of the OS itself, i.e. some aspect of the linux kernel vs. however we presume MS does it in their latest OS. There are no studies out there that say "Linux is faster than MS" or vice-versa. If there are, they should be ignored or at least heavily scrutinized.

To compare some applications' startup times with arguments about which OS is 'faster' is a classic straw-man argument that I despise seeing. The reason why the /. crowd says these kinds of things in their arguments is because they READ THEM HERE!!!

Anyone who thinks less of people who defend Linux should really think about why the Linux community is so defensive in the first place. They have to put up with arguments like this that should be ignored. It's hard to not get upset and defensive when you are presented with one of the most unethical forms of debate tactics ever created.

So far, I haven't seen a single post stating that OpenOffice is faster than MSOffice. All I've seen is people defending their opinion that the difference is not nearly as drastic as the article made it out to seem. In fact, I've seen several posts explaining how to make OO start up faster. It is not obvious that the author of the article has even attempted to try to figure out if there's anything in his control he could have done to level the playing field.

If you want a real argument to debate, here's one: "OO is better than MSOffice because it IS bloated."
If you can provide a non-straw-man argument to support this, whether you agree with it or not, then we forgive you.

IT CAN'T BE OPEN OFFICE!

[mister_llah]

Oh no! It *can't* possibly be OpenOffice! Nooo!

It's gotta be the processor speed, Windows, or *somehow* related to being Microsoft's fault!

God forbid everything Open Source isn't perfect...

===

Look, gang, step one to correcting a problem is realizing there is one. Realize it, fix it, move on. Improve.

Everyone clamouring for reasons why OpenOffice isn't bloated, I'm not saying you are wrong... I am just going to say this (stiffling urge to say "You are not right" even though it would be funny) ...

Even *if* it isn't bloated, wouldn't optimizing the speed even further be a benefit either way?

===

Sure, the article is lacking a bit in the specific test methodology department... but just because of that, don't assume that the guy is automatically wrong.

Just my two cents.

cross-over office

[Anonymous Coward]

To add insult to injury... Cross Over Office Wine runs MS Office under linux faster than OOo runs under linux natively! It is insaine! That is why I have quit using OOo and I now run MS Office via wine.

There is no reason to run OOo except for the fact that you either:
Don't want to pay for the MS Office + CWO Wine
or
You don't want to pirate MS Office + CWO Wine

Excel is smarter

[Mustang Matt]

Excel is the piece of Office that I just can't quit using yet. (Consequently, I do like OO Writer better than MS Word. Especially when it comes to nested tables.)

Excel allows for cool tricks like pasting the following into a text file and naming it .xls:

testtest2
test3test4

If you try to open that same xls in openoffice it opens it in writer.

Silly article

[Javaman59]

It's Microsoft software which is bloated and buggy. Everyone knows that. Always was true, always will be true. Microsoft couldn't code their way out of a paper bag. I use OO all the time, and I know that it's faster than Office...or that it doesn't matter..or that it's because Microsoft screwed Netscape, or it's because I don't understand Open Source, or something....

Re:Children, grow up and admit that OSS isn't perf

[NutscrapeSucks]

Two reasons near the top are that Windows is slower and more bloated. These reasons are sited often and are part of the OSS mantra.

I would say this *was* part of the OSS mantra, but was quiety dropped for obvious reasons when OSS desktops got to the same level of functionality as Windows.

To compound matters, they couldn't agree on API standards, essentially forcing users to keep three or four entirely unique frameworks in core to accomplish the same objective as Win32. Thus bloat went from an advocacy point to something that is nearly an order of magnitude worse on the OSS desktop.

The current OSS mantra is "security", but once again it looks like this talking point might turn around and bite them in the ass.

Re:"Essentially" the same data? MOD PARENT DOWN

[Anonymous Coward]

Because spending money on hardware is cheaper than spending money on software.

Consider "man-hours"

[Anonymous Coward]

Yes, SO has been around a long time, going back to the heyday of OS/2 and German corporate needs for an office/productivity software solution.

Nevertheless, there is no question but that M$ has about a 1000-to-1 advantage in "man hours" spent on its Office over OO.

Although open source development should alter this ratio substantially over time.

What bullshit.

[twitter]

Troll posts have really gone downhill here. It would be better for you to direct your name calling at a real post rather than a strawman. Can't you at least use a sock puppet account to create the strawman? Now, let's get down to technical issues, shall we?

It's time for you to have egg on your face, admit it, and take it like an adult. And then the next thing you need to do is stop wasting your time and fix the problem.

What egg? Gnumeric and Kword run circles around both OO and M$ Office. As a free software fan, I have my choice of applications. Unlike 20 years ago, that's not the case on Microsoft platforms. The primary reason to run OO is to open Microsoft's horrid and bloated file formats with an editor that has good html and pdf export. When I want to write a paper or manipulate data the free software world has more choices that work better than Bill Gate's best wet dream. The problem is solved because it never existed.

I call bullshit on ZDNet. That Wintel rag never gets free software stories right and often gets them wrong deliberately. It's not worth reading and I could care less what gets published there. OO may be big and bloated next to other software but Microsoft is the prize hog and always will be.

In this case, it goes against personal experience. I can and do run OO on systems with resources Microsoft's latest, the now five year old XP, won't even install on. When I need a slide show, I pop open OO presenter on ... a Pentium 2 with less than 256MB of RAM. Try that with M$ crap-ola. Sure, you might be able to create an 80MB monster piece with software that Microsoft has abandon support for. Good luck getting it off that machine on a network, ha ha. Will your new copy of Office render it correctly?

An extreme example of fat new software on skinny old hardware was posted onto my LUG recently. Check out OO running on a 133MHz Pentium laptop with 70MB of RAM:

Holy Shit! That's crazy. But far from 200MB for OO alone, the combined six or seven big applications took 120MB of swap. The nutcase ran OO, Abiword, Kword, Kontact, Konqueror, GIMP and other software at the same time on separate virtual desktops. XP would be a fragging, blue screened disaster if someone tried to run M$ Office, Outlook, Photoshop and IE all at the same time, say nothing of trying to load up a few other extras.

When someone does that with XP or Vista and M$ Office, I'll believe that they are no less bloated than OO and Linux. Right now, I believe, XP won't install on less than a P3.

Re:What bullshit.

[Anonymous Coward]

Can't you at least use a sock puppet account to create the strawman?

You mean like yours?

database tutorial

[Hanzie]

Given that: I have to admit guilt. Does anybody know of a really good intro to Base (oo.o's database). I need database functions, but every time I try to get started, I end up floundering.

If anybody knows where I can find an excellent book/ tutorial/ whatever so I can learn Base, I promise to stop making bloated spreadsheets. (MS Access tutorials are, however, out of the question.)

Thanks very much, any who respond helpfully.

hanzie.

Even with <100 points

[achurch]

I've got a finance spreadsheet with about a dozen graphs, each with 30-60 points. Excel can recalculate them all almost instantly (< 1 second). OOo takes about 20 seconds. That's 20 seconds every time I hit Enter. Argh.

Huh?

[Seraphim_72]

Who the fsck modded this "insightful"?

with a proprietary code base

Huh? Go get the source. Proprietary? To whom?

an open-sourced proprietary product.

What? Perhaps if you type in a less MS shill tone I could see what in the heck you are after. OOo is open source, there is no propriety to it, grab the source, fork it all you want. Is SUN a major contributer? Yup. Is Star Office OS? Nope. Does OOo get all the cool stuff from Star Office as OS? Yup.

Really - what is your beef with the licence? OOo *IS* OS ... there is no "Proprietary" to it.

Sera

Re:Huh?

[alienw]

You completely misunderstood the point I was making. Here, I'll make it again, maybe it will be more clear

Huh? Go get the source. Proprietary? To whom?

OpenOffice used to be a proprietary product called StarOffice, made by a company called StarDivision. The codebase is at least 10 years old. Yes, it's open source now, but almost all of the code is the old StarDivision code. This type of code (monolithic and self-contained) does not translate well into open-source development. Nobody wants to dig around 3 gigs of source code (yes, that's how much it takes up). The only people really working on the code are Sun engineers.

If the OpenOffice codebase was developed as an open-source project, things would have been done much differently. The main problem with commercial software is that it does not build on the work of others. OpenOffice has everything built-in including its own windowing system, graphics system, database, IPC mechanisms, a component framework, and so on. It is hideously over-engineered. This yields a huge codebase, a huge memory footprint, and poor maintainability. These types of projects generally attract little or no community participation, and ultimately fail as open-source.

There is no evidence supporting your argument.

[WebCowboy]

This is one area where Open Source has its weakness.

Cutting down and optimizing existing code is not nearly as glorious as adding new features.

Exactly how did you arrive at that conclusion, by pulling it out of your butt after reading this one article? There is absolutely no evidence to indicate that open source promotes feature-creep and bloat. In fact the opposite could be true.

Try to compare OO.o to previous versions of itself rather than just MS Office. On the same hardware you'll find OO.o starts faster and consumes less resources than previous versions. I've NEVER seen the same trend in MS Office or ANY Microsoft products--in fact major upgrades of ALL MS software have consumed more hard drive space and had the same or higher recommended system requirements. Yes, OO.o probably does still not compare well for efficiency compared to MS Office, but with each new release the gap closes. I'm betting that the next version of OO.o will be a bit more efficient, while the next release of MS Office will be less efficient.

Your assumption that adding features is the most "glorious" task and thus receives the most attention from F/OSS develoers is also somewhat flawed. Efficiency and clever solutions tend to be the greatest source of pride in terms of engineering, and I find that open source developers usually take a more engineering-like approach to their creations than commercial developers, where marketing is the biggest driving force. It seems to me that the FOSS culture is influenced by the open culture demonstrated by early PC hardware hackers like the original homebrew club. Anyone could (and did) make their creations do more. The ones who did more with less were the ones who had all the glory. Steve Wozniak's Apple I and II designs didn't get accolades because they had the most features--even in the late 70s you could get Altair/S100 bus video cards to provide "modern" I/O, colour graphics, sound, etc. Woz was heralded for providing those desirable features in elegant, very clever and hackable designs that used far less components and cost significantly less than competing designs.

As FOSS projects become more "free" it seems they become more efficient and more modular. The Mozilla foundation is a good example--they inherited a bloated, monolithic and unmaintainable closed-source project in Netscape Communicator. It didn't take the artitechts of Mozilla long to decide to scrap it all and make the code more modular and maintainalbe...and evenually that thinking extended from the architeture to the end applications, when a set of more lightweight applications (Firefox, Thunderbird, Sunbird) has taken the spotlight over the big, feature-laden traditional Mozilla suite.

The OO.o project has a history similar to the Mozilla project, but the evolution of the project has folowed a somewhat different path. The OO.o project has roots in a closed-source office suite from Germany. Sun acquired it and subsequently opened the project. I personally dislike the architecture of OO.o as it is too monolithic for my tastes. My opinion may be biased by my first impressions of the first releases of OO.o...if left a bad taste in my mouth like when I tried to used Microsoft Works (you know, the taste you get when you throw up in your mouth a little bit...). I still hold out hope that those involved in the development of OO.o make a push towards "breaking things up" so that Writer, Calc, Impress, Database can be downloaded and used independently of each other like Firefox and Thunderbird. I think that because Sun is the sponsoring organisation and has such an influence in steeering the project that this will be tough though. I think the evapouration of Netscape corporation was really the best thing to happen to Mozilla--it has meant that the foundation has steered the developemnt of the Netscape product, where with OO.o it still seems to be the wrong way around--like StarOffice steers OO.o where Sun wants too much.

In any case, there is another benefit to F/OSS:

Re:There is no evidence supporting your argument.

[Agarax]

That, friends, is the rant of a zealot.

See what I found in the newsgroups comp.office.app

[gummyb34r]

~~~~

From: Guy de'Geek
Newsgroups: comp.office.app
Subject: What would you like to see most in openoffice?
Summary: small poll for my new office suite
Date: 25 Oct 2005 20:57:08 GMT

Hello everybody out there using OpenOffice -

I'm doing a (free) office suite (just a hobby, won't be big and
professional like OOo) for 586(686) stones. This has been brewing
since april, and is starting to get ready. I'd like any feedback on
things people like/dislike in openoffice, as my suite resembles it somewhat
(same layout of the GUI (due to practical reasons) among other things).

I've currently done numberscruncher and wordscruncher apps, and things seem to work.
This implies that I'll get something practical within a few months, and
I'd like to know what features most people would want. Any suggestions
are welcome, but I won't promise I'll implement them :-)

PS. Yes - it's free of any openoffice code, and it is fast,
multi-threaded and not greedy to resources. It has descent charting too.

~~~~~

So there is hope...

OpenOffice saved me MUCHO time today over Excel

[CoolHnd30]

I had a user with a corrupt Excel spreadsheet. She wanted me to dig up a tape backup from March to restore. Instead I opened it in OpenOffice, and saved it back to Excel format, and voila, it mirculously opened in Excel again. So, in this instance, at least, OOo was much faster. :)

Re:OpenOffice saved me MUCHO time today over Excel

[TheBAFH]

I had exactly the same experience, with corrupted Excel file. This is a fine example about why monoculture is bad.

Re:Children, grow up and admit that OSS isn't perf

[Dcnjoe60]

I'm not sure how valid memory usage for empty documents really is. For example, let's say that there are two identical programs, coded exactly the same way except that one of the two programs allocates a lot of buffer space up front to keep from having to do it on the fly (since memory allocation is usually cpu intensive). Which of the two programs is the better one? Of course, the answer might be different if one is trying to run on 128MB vs 2GB.

Don't get me wrong. OO has lots of room for improvement, but I'm not sure that empty document memory usage is a valid measure. For one thing. All of those Windows DLLs that Office requires but are part of the OS don't get counted in the memory footprint. Now it is true that they are part of the OS but technically, they should be counted. OO doesn't have that luxury -- all of it's memory footprint is displayed, because it can't rely on the Windows DLLs and remain multi-platform. If one day, Microsoft is successful in integrating .Net with the OS, then we'll here the same kind of arguments -- "Look how much memory a java app takes, when C# uses next to none!"

All of that said, OO does suffer from having a monolithic architecture. You shouldn't have to have Writer and Impress loaded just to use Calc.

But the OEM MS (Basic) Office Suite is cheaper.

[WoTG]

True. But you can't by Excel "OEM" (at least I've never seen it). An OEM copy of the basic MS Office Suite (Excel, Word, Outlook) from Dell is ~ $150. From your link, Excel itself is $229. I think I'll take the suite.

Re:Children, grow up and admit that OSS isn't perf

[Dcnjoe60]

I don't think it's just FOSS advocates that use load times and memory footprints but instead most people do. It's something that on the surface appears to be objective, but in reality is not. Businesses don't care about either. When they evaluate applications, it's based on functionality, productivity, cost, etc. Memory footprints only factor in if additional memory is needed, thus driving up the cost. The only people that really care about load times and memory footprints are the technical people. Average Joe user, isn't even aware of it, or if he is, doesn't understand it.

Case in point. Most new Linux users get all concerned when they check memory usage and see that they are near maximum. They don't understand that the OS is using all available RAM for it's own processes until something in user space requires it.

FOSS may report or quote things like load times and memory footprints, but if so, it's only because people expect them to. FOSS never even mentions any of those kinds of things when describing what FOSS actually is.

In reality, comparing load times and memory footprints is about as meaningful as using MIPS to compare mainframe systems. What's the joke about MIPS? MIPS=Meaningless Indicator of Processing Speed. Likewise, comparing memory footprints and load times is meaningless for the average user.

Why JAVA?

[Anonymous Coward]

Why use a db engine written in JAVA when there's already SQLite?

Re:Children, grow up and admit that OSS isn't perf

[nathanh]

So I find it incredibly ironic that now that the shoe is on the other foot, the tables are turned, etc., that these very same people are dismissing "bloated" and "slow" as unimportant.

Are the "very same people" saying both things? I must admit I don't keep a database of what everybody said, so if you could post an example of a single person saying both contradictory things then that would be super.

Re:Children, grow up and admit that OSS isn't perf

[bheer]

You, sir, ought to get a prize for showing exceptional idiocy even by /. standards.

All of those Windows DLLs that Office requires but are part of the OS don't get counted in the memory footprint.

Wrong. Wrong. Wrong. Which DLLs are these? GDI? User? AdvApi? Kernel? Here's a clue: OO uses them as well. So does every other Windows app (that includes java.exe and javaw.exe)

Now it is true that they are part of the OS but technically, they should be counted.

They aren't counted in the working set, but they are counted in the virtual size (which is a measure of how much of a process' address space is used up). However, virtual size (for the purposes of this discussion) doesn't matter because its amortized across apps. The OS' shell (explorer.exe) already loads most of these DLLs. So any apps that links against these DLLs will 'run faster' because most of its stuff is already loaded.

Unless of course they insist on loading LOTS of their own dinky libraries at startup time. Like Seamonkey pre 1.7. Or OO.o. One reason Firefox got popular was because its developers actually had a clue about writing good Windows programs and took advantage of almost every trick in the book (I use XP and Ubuntu Hoary and Firefox always starts faster on XP).

Translation: code targeting a particular OS' features and environment will always run faster than generic crossplatform code. Big surprise.

If one day, Microsoft is successful in integrating .Net with the OS, then we'll here the same kind of arguments -- "Look how much memory a java app takes, when C# uses next to none!"

Wtf? Talk about clueless paranoia. Please. Someone give a the parent a Purple Brain for stupidity.

Slow with same data?

[john666seven]

Yes, Excel does run slightly faster than Calc--who cares? It is easer to use. At 2'5 GHz you don't notice. Personally, I only use Excel for training to use Calc better (it works) since there is lots more information on Excel. By the way, I have found that Word 2003 as part of the full office suite is the most unstable word processor on the face of the planet--it has destroyed it's own documents for me. do you want to type 27 pages all over again because Word had a problem while you had the document open??? NO,I didn't think so....

Re:Children, grow up and admit that OSS isn't perf

[AaronLawrence]

The only people that really care about load times and memory footprints are the technical people. Average Joe user, isn't even aware of it, or if he is, doesn't understand it.

Most people won't care about memory footprint (so long as it doesn't cause swapping, a big if), but every person notices slowness. There is a reason that UI guidelines suggest responding in 0.1seconds to user's actions, and that should ideally apply to loading programs as well.

OOo is much faster than opening MSO through Citrix

[KayakFun]

I work in a mixed Sun Solaris/PC WinXP environment. To allow Sun users to view/edit Office (MSO) documents, we had to start up Citrix, login with username and our ultrasecure (=long and dificult) password, wait at least 30 seconds, and then see Citrix open the document. Major PITA.

With OpenOffice (OOo) being a Solaris-friendly application (both are made by Sun), we can soon open those Office documents a lot faster, and save a Citrix license, an Office license, and a bunch of heavy PC servers running Citrix. Native applications rule!

But the story does not end there. Once all Sun users use OOo, eventually someone will break a document because of incompatibilities between OOo and MSO in handling the Office format. Therefor we are now already looking into making (OASIS) OpenDocument (ODF) our default format for certain development documentation that is primarily authored by Sun users. IT must then roll out OOo on all PCs to be able to view/review/edit ODF (MS refuses to add support for ODF).

After that it's a small step to a full switch to ODF, especially since the XML is totally open and preferable over MSO-12 (our Sun users must also view the XML). By then the beancounter will come in for the kill: why spend MSO license money for 5000+ employees when you can do with a free and 99% compatible alternative, that even kills the necessity for Acrobat Elements (PDF generator for 100+ user companies) and a lot of Citrix licenses and hardware.

The trend in IT is 'cutting cost'. So we will probably end up with the cheapest hardware (PC) with the cheapest OS (Linux), with the cheapest office suite (OOo), browser (Firefox), email client (Thunderbird), HTML editor (Nvu), graphics suite (GIMP), graphics conversion (ImageMagick), and so on.

Guess what: most of that software already runs in our company on Solaris and on WinXP today. Switching to Linux PCs is a breeze because users are already got used to the applications over the past years one by one. No costly retraining. Some users may not even notice the difference between SuSE Linux and WinXP, other than that the applications are now grouped together by type in the Start menu, rather than by manufacturer.

It's faster than I type, so fine!

[Anonymous Coward]

I don't see the problem. Yes it takes a few seconds to load a typical file I use, but I can't type fast enough to race ahead of it. When I'm typing as fast as I type (which is pretty fast) my system CPU meter in gkrellm barely moves. OO may be "slow", but the human driving it is slower and is the limiting factor for how fast text can be entered.

OO for me is free. It works. It was ahead of Microsoft Excel for putting formatting inside spreadsheet cells, so for drawing sound cue sheets for theatre with formatted instructions for the operator it wins hands down. I couldn't do things like make "Be sure to patch to group 2" bold in Microsoft Office.

I have a friend who's been on an old MS Office and is having difficulty moving her documents to new versions. Open Office is a winner here! I'll have to see how the migration goes.

I'm running on a 600MHz Mini-ITX M6000 with 512M RAM. I don't know how Windows XP's memory management compares with Linux. Older versions of Windows were pretty poor. I'm running the Debian OpenOffice packages with the Gtk/Gnome support so should be getting some help from shared libraries. I remember the old Debian OpenOffice installer used to prelink, though I don't think this one does. Program load time is not too bad though.

Er, so what do you lose?

[pjc50]

If it can be disabled without losing functionality, what's it there for?

(I can't find the Java option in my OOo 1.1.3)

Price Price Price

[erraticorbit]

My Hyundia Excel does not out perform my neighbours ferrari. Yes it is true but what do you expect when you compare the price

Alas, puns don't get the attention they deserve

[kylef]

I realize that it didn't get moderated as such, but *I* thought it was a funny post... :-)

Re:shoe on other foot this time? weird.

[emilper]

last time I used MSWord willingly (true, it was in '99, on Win95 and it was version 6), it crashed after it mangled my file and *saved* it. I could not open that file again using Word or another text processor, and had to edit it in MSEdit to delete the binary crap and unshuffle the text ... managed to recover less than half ... OOffice still has to beat this ...

Re:shoe on other foot this time? weird.

[alienw]

If I only wanted to process text, I would use vim or wordpad. Maybe you only use your word processor for 9th grade English assignments, but most normal people use them for things like research papers and reports and such. These often have more figures and equations than text.

Useful Memory Test

[KayosIII]

Ok I just did an exmap test on OOo Writer - I am getting ~50mb of effective memory usage. That is still the heaviest thing on my system apart from (x.org)...

I haven't actually got any document open (its a blank page) and I have disabled Java. I would love to test Microsoft Office but I have neither the money to waste nor a system to run it on :p

Astroturf

[SgtChaireBourne]

MS marketing includes a lot of freelancers:

'Many are specially trained, sometimes at corporate headquarters, Gossett said, as in the case with Microsoft. They are expected to devote about 10 to 15 hours a week talking up the products to friends, securing corporate sponsorship of campus events and lobbying student newspaper reporters to mention products in articles. They also must plaster bulletin boards with posters and chalk sidewalks -- tactics known as "guerrilla marketing," which, marketing firms acknowledge, intentionally skirt the boundaries of campus rules.'

Now how exactly is M$ not like a MLM anymore?

The special training at corporate headquarters is probably one of the reasons there is sometimes a hiatus and it'll go a few days without a peep in defense of M$. They'll also attack, usually with logical fallacies (e.g. ad hominem), any criticism or even critique of products or initiatives that are being launched. (e.g. right now MS SQL.) Read carefully the next attacks from MS fanbois and see that they usually change the topic or go into name calling.

Anyway, it's not a surprise to see them go after OOo and less so for OpenDocument. Both cut into their MS Office revenue. OpenDocument cuts off the lock-in at the file format level, removing dependence on MS for continued use of the documents.

hum... indecision

[Anonymous Coward]

hum, let me see....... In my desktop I can get
$0 OpenOffice.org loading slower,doin anything Office does,supporting OpenDocument (my files will be good FOREVER)
$250-300 Office loading faster, supporting closed-formats (when I'll decide to change office, I'll have to pray MS...)

hum... difficult choice.... but I think I'll use all those bucks in another way:
$50 donation to openoffice.org
$150 new cpu/motherboard to my old system. Now not only OpenOffice is goin nearly as fast as Office was goin on the older
                    system, but all other apps are faster......
$50-100 Cinema,DVD, meeting girlfriend.....