I second this. I've been using Bibble 4.1 on FC2 since shortly after it was released. No stability problems, and it is *way* faster than GIMP for similar edits.
While Bibble can't do layer operations, dodge/burn, and cloning like GIMP can, it can rotate, crop, sharpen, adjust levels, saturation, and EV. If you depend more on your camera for composition than your editor, Bibble will get you quite far in many cases.
The downfall of Bibble on Linux is that there isn't any good way to profile your monitor. Thinkgs like OptiCal (calibration puck) just aren't supported on Linux. So, even though Bibble supports ICC profile transformations, you can't get an accurate profile for your monitor under a Linux video driver. That pretty much ends the game before it begins.
I think usability of the GIMP is an excuse when it's in the context of interfaces, but anyone who thinks adjust levels in an 8bit color space is adequate has never seen an image adjusted in a 16 bit space. No more JPEG jaggies when I use Bibble.
GIMP is a fantastic tool, and I'd love nothing more than to use it exclusively. But until Xorg supports integrated color management, there's no point in GIMP having 16 precision. It's like trying to to precision machining with a tape measure.
There are a LOT of sites using SVM in very real tasks. It also (as of S9) has some excellent features which compete heavily with their Veritas equivalents. Things like logical volumes, snapshots, and increased performance. With S10 and ZFS there are even fewer scenarios which would require Veritas.
Even before SVM (when it was SDS) it was used heavily because so many sysadmins are in therapy for mental damage incurred while unencapsulating their Veritas root volumes. For smaller systems, using LVM is cheaper and less prone to error in a failure scenario. For larger systems, using SVM for OS volumes and (if necessary) VxVM for the disk farm is quite common.
I think all the griping about the UI is pointless at this moment. It's like beggong for a square steering wheel before your car's transmission has 2nd gear implemented.
GIMP has roadmapped 16 bit editing, CMYK, and color management. These three things are ubelievably critical to pulling over larger numbers of digital photographers. The artistic world is in pretty good shape with GIMP, the web graphics world is doing well with GIMP, and the digital photographer has no choice but to leave Linux, or purchase Bibble. Not that Bibble isn't an excellent product, but it's never healthy to have NO competition.
I also believe that GIMP's implementation of color management could be the spark that wakes up companies that produce colorimeters, and perhaps kicks off system-wide color management through X, and support for ICC profiles in CUPS. I can only dream how cool it will be...
It's not safe to assume that clustering on its own will accomplish this. Most Enterprise cluster environments add significant administrative complexity. Given that human error is responsible for vastly more downtime than actual hardware failures this is significant.
The most effective means of improving availability is through automation of administrative tasks. For clustering to have a long lasting impact you first need a sufficiently mature IT practice. This alone can get you to a very significant level of "9's".
The second thing is adequate disk redundancy, and the third is environmentals. By this point you're at 3-4 9's consistantly and clustering is gravy.
I agree that the tools are (LTP) are rock-solid and critical to Linux' growth. I also agree that Linux is ready for prime-time, and aknowledge IBM's significant contributions to advancing Linux in the Enterprise. I even agree with LTP's purpose not being to proove anything.
My comment was specifically geared towards the article - not the tools, or the product of Linux. I think the article was very soft on details which is unfortunate given the talents IBM has working on Linux. The article seems to have its purpose rooted in adding credibility to Linux reliability, but I don't view the test as showing that. I believe you need a larger sample (test period) to make a conclusive argument about reliability, and a more wide range of tests.
I work as an availability engineer (studying reasons for downtime and finding ways around them) and the trends I see often take a year to show themselves. We won't even perform an analysis until we have 90 days of data, and that's rarely enough for a trend. An OS' reliability is based on much more than running a load of 90+% for X days. Very few systems actually run with that kind of sustained load in a production environment. Reliabiliyt is also heavily based on it's fault isolation, and resiliance to the unexptected (all of which Linux excells at IMHO).
I just don't think that this test did what it set out to do, although I'm sure marketing departments think otherwise. I would have covered various types of fault injection, and liklihood of human error for common recovery processes as well. I would also have had 90 days as my low end, not 30 days. Finally, I would have put these comparisons next to other options like Windows and a UNIX or two. I don't think the results would have changed, but I do think the results would be significantly more credible in prooving its capabilities.
I just finished reading the article, and I'm disappointed. It seems to me that the purpose of this article is to provide yet another "Enterprise-ready" stamp of approval for Linux. Now, let me be the first to say that I'm all for Linux in the Enterprise, but this article isn't the way to do it.
In the Enterprise world reliability equates to availability. This test used high loads over 30 and 60 days. To be Enterprise capable, in my mind, means 2-nines, or 99.0% availability. That's 3.65 days/year or 01:41:00 per week. To be impressive (and worthy of an article) I'd be looking for a minimum of 4 nines (0:52:30/year, 1 minute per week). Most mature sites I work with are closer to the higher 3-nines range on a consistant basis using a commercial *NIX.
What I'm getting at is that to have a real test give you a sample size that you can trust, 90 days is the bare minimum. I would have expected a 12 month study, and a much better explanation of Linux ability to deal with HW errors. IE - inject disk failures and scenarios which test fault compartmentalization. Doesn't matter what hardware you're running on - a good solar storm can play all kinds of havoc in a data center.
I'm disappointed that IBM would consider this anything more than a science project. Knowing the resources and talent they have at their disposal I was expecting more.
I'm wondering if the trend of having more workers telecommute recently is helping to keep the pricing faith alive for broadband providers. It's easuer for someone to agree to $40 or $50 a month if they're expensing it...
I would LOVE to find a way to convince myself that this just won't go anywhere for them. Unfortunately, I think it might work - take script kiddies as an example. They don't get the art of hacking, they just want to be able to brag about what they did (not how they did it). It's soul-less.
I think that's how a lot of computing is today. How it looks is very important to people who've never been exposed to the alternatives. I used to actually like Windows (I know - not good to say on/) until I went to college and hung out with a dude who showed me slackware. Until you know there's another road, you never notice the bumps in the one you're on.
As much as Linux is making impressive progress, I don't think that road has been mapped yet for the masses. I hope I'm wrong though!
Maybe I'm missing something but it sounds like something that could be remedied by using page styles. Set up a page style with the numbering you want and assign it as needed.
I had different numbering formats I needed to use and just about pulled out my hair trying to find a Microsoft way to do it. Once I found a good book on SO/OO I fifugred out page styles and problem was solved.
Might be a different issue, but sure sounds the same.
Actually, there's more to it than that. I wish I had a better handle on the internals, but I have some documents in OO (SO actually, but same) which I can only PDF in the newer release. If I use an older release, or ps2pdf, the fonts come out jagged. Tried the same thing on Adobe's test drive on-line converter and they came out the same way - ugly. SO7, OO1.1 both convert nicely though. Very bizzarre.
No idea why some things work and others don't, but all PDF converters are not created equal. It would be interesting if someone who knows about this stuff could explain it.
Of course, M$ Office doesn't include Project Management software either... It's an EXPENSIVE add-on. Not sure how Mr. Project stands up, but RedHat includes it in its distro.
Reliability comes to mind, but I will concur that high levels of reliability in a workstation aren't needed by everyone. I have an Ultra 60 (2x450mhz/1gb) right next to an Athlon 2200 with 1.5gb. While I agree that the Wintel machine is faster, I disagree with the level of sensationalism you attached. I would also suggest that if you are getting 1mb / second throughput on a Sparc box you have some kind of a problem. Have you checked/var/adm/messages lately?
Here's another point to consider... Your 286 which you claim is fast as the Sparc box can't run Windows 2000. My Sparc 20 lab box is running Solaris 9 comfortably.
The list goes on, but your note is such an obvious troll I'll stop here.
SO versions 7 exported some ugly PDFs if you wewren't on a mainstream font, but the PDF engine in SO 7 is a completely new one with 3 levels of quality (Draft, Print, Press). I can't explain it, but when I demo'd Adobe's latest release it still produced jagged edges on one of my typefaces when rendered to PDF, but SO produced it cleanly.
Not sure if that's what you referred to by ugly, but if so, it's fixed. If you were talking about the interface in general then what the heck do you call attractive??
Is Sun really on the ropes, or is that what the analysts are saying because they don't "get" Sun's Linux strategy yet?
Take a long hard look at the R&D budget and the cash in the bank and I think you'll find that Sun is alive and well - even in comparison to most of its competitors.
Star Office is not very similar to Open Office at all, sun kept the best parts to themselves (database app) so why are they seen to be *cooler* to open source zealots then other perfectly good office sweets?
This is flat out wrong. I use SO on my work machine and OO on my home machine. The only significant difference I encounter is in the standard fonts bundled. More information on the differences can be found in the OO FAQ.
You are correct in that the database component is only included in the StarOffice version, but is that really so bad? If it's such a problem than try this command: "rpm -ivh postgresql". It's not like Adabas is a MSAccess killer at this point anyway.
Let's put this in perspective: Sun bought StarDivision, and opensourced it immediately - that was pretty nice in my book. If you remember, SO 5.2 was entirely free until the business customers of Sun expressed doubt in the longetivity of the SO business model and requested that it be for a fee. Sun complied, and left OO in place as a free alternative for those of us less concerned about the business model. Oh, and SO was free for educational use.
I have trouble believing that someone who claims they are not very similar has spent much time working with them, but I suppose it is possible. It always bothers me to see FUD in the open source world; Don't let the dark-side cloud your judgement young Jedi.
I think that engineering software for reliability does not cost any more, and, in fact probably costs less, than engineering it for unnecessary whiz-bang features.
I agree... I'm just not sure that the marketing organizations which are hungry for the "value add" would support that slim-down approach.
I also see where you are coming from in terms of calling software a product. With the right definitions in place your Honda analogy makes sense, but it would probably cause heart failure for most software businesses to think that way in our current litigation-crazy society.
The big point in your statement, which I think most companies wouldn't go along with, is that it costs less to design for reliability. Let me say that *I* agree with you. My supporting evidence for doubting business follows this line of thought is the emphasis on quick turnaround and time to market vs. quality. Short term gratification is what drives most businesses because they aim to please financial analysts more than consumers.
Look at the Ada programming language... It was designed to facilitate more reliable programming of large scale projects. Yert, most programmers don't like it because its "a pain in the a**". Yet, if you look at statistics, a cleanly compiled Ada program is much less likely to have bugs than a cleanly compiled C program. I know there's more to it than that, but my point is that getting engineers and businesses to look at the long term is going to be very difficult in today's hemorrhaging tech-sector.
I'm not holding my breath waiting for that to happen.
Look at a laser vision correction system. It's got extremely complex software and you don't hear a lot about those crashing while performing surgery.
You make some good points here. I certainly don't have your experience - my only "stick time" was a college project on a RT-Vax, but I'm not entirely ready to abandon my thinking. My (revised) thought would be that most RTOS environments used for process control on things like laser vision correction and Boeing 747s were developed to much higher standards of both QA and architecture than an average PC operating system (like Windows). So, in some ways that prooves your point.
Because an RTOS will often have a human life at stake it has to bedesigned to higher standards. All that extra testing, design, and higher-skilled engineering time costs significantly more $$, and may be enough to make the costs of an OS prohibitive to a commodity PC. I would imagine that an extra $100 per PC license would have a big impact on how many Dells came loaded with Windows vs. Linux. The mere fact that an OS may have lives at stake creates a different programming atmosphere - one of responsibility than not all software engineers are up to.
Another point I would suggest is that an RTOS isn't doing as much as a regular OS. It would typically be stripped down to the essentials for its function. In contrast a modern OS is doing all kinds of things at once with combinations of software not anticipated by its designers, thus the integration is a crapshoot.
I think this lack of control is probably the biggest deterrant to OS manufacturers; If they were the only ones designing software for their OS they probably *could* warrant it. Unfortunately, who's to say what quality went into an arbitrary package loaded on their OS? You would never see a third party screen saver driver for a lasik machine (ok, bad example, but hopefully you know what I mean).
I do agree wholeheartedly that the average consumer has allowed themselves to tolerate much more sloppy coding than they should - we should hold vendors to higher standards (and hey, I work for one of those vendors!). The most obvious example of this is how many people tolerate needing to reboot their machine at least 1/day. Some people are amazed that my Linux workstation goes 100+ days without a reboot, which is sad. But I think it will take a few years of cultural inertia breaking before we could see an OS which could be warranted as a product. It would take a change to the mindset of engineers, engineering managers, marketing (who drives feature roadmaps) and executive management. That's no small feat, although definitely possible.
I think headcount makes a lot more sense than the more common per-CPU licensing we see on some enterprise software. Remember that the per desk fee includes Professional Svcs to make the transition, 60 hours/week of support, training. Also note that its per desk, so the gardeners mentioned in an ewarlier post wouldn't be counted;)
Compare all of that priced out to what it costs to run oracle on a 20 CPU server, then imaging you have a DR site, a cluster with at least one more node... You get the picture. For a lot of sites this will be a huge reduction in cost.
It will be interesting to see if this experiment works. I agree with you - it may be viable for OSS as well.
Seems like there's a pretty big difference between something that is mechanically engineered and something that is software engineered. The term "product" IMHO applies to something that can be mechanically engineered (constraining myself to the Chrysler vs. software here). If you try to apply it to SW engineering I don't see how given today's technology it could be warranted as a product.
Chrysler doesn't have to depend on compilers and virtual machines and underlying hardware in the same way that software does...
I think you have a valid point lurking in there, but the analogy is lame.
I don't think I implied in any way that I'm a God - I only said that I do a lot of presos so that you'd know I wasn't a casual user. I certainly made no implication of your abilities, although your response certainly does. I'd like to hear more "facts"... Unfortunately your response doesn't include anything but conjecture.
You mention that "Corporate uses PowerPoint"... I wokr for a very large, very global company with all the same corporate red tape other large companies are hindered by. We use SO as our office platform and interperate with our MS Office using customer base with no significant problems. Many large goverments (talk about corporate red tape!) as well as educational institutions are also getting the message recently.
One of the things my team does is organize global customer forums where the customers present their best practices. Most of these presentations are in Power Point, and with few exceptions (mostly due to font translations) they load up in SO with no significant issues.
The presos which we deliver are authored in SO7, so obviously need no conversion. If you have specific areas which you can point out in which PP is superior to SO I would be interested in hearing about it. My observation is that very few of those areas amount to more than a a "cooler way to embed a MIDI file into my slide". I'm interested in learning about other areas as I haven't spent as much time with PP.
As for the spreadsheet, the only area I'm aware of SO falling significantly behind Excel is in 3d surface plotting. When we were graphing performance scenarios for enterprise storage we ended up using Excel for twiddling and GNUPlot for the automated suites because they could do it and SO couldn't (see, I can be objective on the negatives as well!)
Not sure where you picked up the negative vibe, but if you can carry a conversation I'm game to learn what you know. The question is, can you open your mind?
I was pretty happy with the functionality of Word 2.0 in terms of features. I'm not sure that pseudo-AI in the form of a talking paper clip has improved my writing skills.
I think the pricing of something like an office suite (which IMHO is a commodity) is insulting. I think M$ would actually benefit more from including it free in Windows as a value-add than charging astronomical fees. It would be yet another notch in their office ubiquity and even further solidify it as a standard. What's more, it would bring Windows 1 step closer to competing with Linux (Linux does so much by virtue of included apps, where Windows out of the box does very little until you install apps).
Of course, I'm perfectly happy watching Linux keep building its lead, so Mr Gates can feel free to disregard my advice:)
I don't agree. "got to have powerpoint" makes no sense in itself.
For most of the presentations I do (and I do a *LOT*) I ignore all the bells and whistles and focus on my content. I put the bells and whistles into the way I convey the information rather than the file itself.
What I end up with a clean presentation file which exports very nicely to PDF where it smaller in size than the source file, and can be read on any machine, including those not powerful enough to efficiently run a complete office suite.
I've said it before, and will say it again. If Powerpoint is needed to make a point, then there are serious communication skills lacking in the presenter.
The other point, which I think you intended but didn't include explicitly, is M$ format problems.
If Joe student bought a copy of Excel 98, and the book ships with a spreadsheet in Excel 2003 that student is probably screwed.
If the book ships with a copy of OO, they (as you suggest) can allow multiple platforms to access that file without worry about format compatibilities.
It's amazing how Apple has been slammed for obsoleting their customers, but M$ doesn't get the same bad press. I have a site I volunteer at which has to maintain licenses for Acess 95 and Access 2000 because they have Access 95 apps which don't work properly in Access 2k and vice-versa. So much for the benefit of backward compatibility.
This is one area which I find frustrtating. For all but the most advanced power users, OO/SO really provides what's needed for business functions.
The only exception is MS Access, however I think a lot of Access apps out there would be just as functional in a spreadsheet since the average user doesn't "get" relational theory, and doesn't program.
When the bastardized M$ XML appears I think we'll have an interesting potential for phenomenon. Historically, (as much as I hate to say it) M$ could do what they wanted without fear of repercussion because the alternatives weren't upo to snuff. Today a company has the very real ability to choose SO/OO as their office platform and take the initiative to ask their sources to use more open formats. This could be the first time customers of Mr. Gates can actually say "no" and choose an alternative.
Does this sound unreasonable? Sun Microsystems uses SO as its office platform (not surprising!). To be fair, I should admit that I'm an employee of Sun, but I think I can make fairly unbiased observations. In the 2 years that I've spent with SO,OO, and all the betas in between, I haven't yet encvountered an insurmountable document interchange issue. I have spent a lot of time writing technical documents, so I'd consider myself in the "power user" camp.
The most common "yeah, but" is presentation macros. My response: If you can't make your point without powerpoint animations & macros, then you don't have a good point. A good presentation *can* be communicated using PDF slides created by the M$ originator if it doesn't convert cleanly into Star Present. What if your customer doesn't have SO? Send 'em a nicely rendered PDF - SO7 has a completely revamped PDF engine which runs circles around SO 6. Better yet, use TEXT where its appropriate, and RTF in other places. If you're really hooked on format abuse, SO7 can even create a Macromedia Flash presentation.
People have come to make 2 mistakes because of features in common office products:
(1) relying on gimmics to communicate rather than relying on mastery of communication skills.
(2) assuming that just because everyone else saves in a proprietary, short lived format, that they need to also.
Given that so many other aspects of the tech industry are moving to value "good enough" technologies, I think we're very close to the tipping point for Office platforms as well.
Slashdot Mirror
I second this. I've been using Bibble 4.1 on FC2 since shortly after it was released. No stability problems, and it is *way* faster than GIMP for similar edits.
While Bibble can't do layer operations, dodge/burn, and cloning like GIMP can, it can rotate, crop, sharpen, adjust levels, saturation, and EV. If you depend more on your camera for composition than your editor, Bibble will get you quite far in many cases.
The downfall of Bibble on Linux is that there isn't any good way to profile your monitor. Thinkgs like OptiCal (calibration puck) just aren't supported on Linux. So, even though Bibble supports ICC profile transformations, you can't get an accurate profile for your monitor under a Linux video driver. That pretty much ends the game before it begins.
I think usability of the GIMP is an excuse when it's in the context of interfaces, but anyone who thinks adjust levels in an 8bit color space is adequate has never seen an image adjusted in a 16 bit space. No more JPEG jaggies when I use Bibble.
GIMP is a fantastic tool, and I'd love nothing more than to use it exclusively. But until Xorg supports integrated color management, there's no point in GIMP having 16 precision. It's like trying to to precision machining with a tape measure.
There are a LOT of sites using SVM in very real tasks. It also (as of S9) has some excellent features which compete heavily with their Veritas equivalents. Things like logical volumes, snapshots, and increased performance. With S10 and ZFS there are even fewer scenarios which would require Veritas.
Even before SVM (when it was SDS) it was used heavily because so many sysadmins are in therapy for mental damage incurred while unencapsulating their Veritas root volumes. For smaller systems, using LVM is cheaper and less prone to error in a failure scenario. For larger systems, using SVM for OS volumes and (if necessary) VxVM for the disk farm is quite common.
I think all the griping about the UI is pointless at this moment. It's like beggong for a square steering wheel before your car's transmission has 2nd gear implemented.
GIMP has roadmapped 16 bit editing, CMYK, and color management. These three things are ubelievably critical to pulling over larger numbers of digital photographers. The artistic world is in pretty good shape with GIMP, the web graphics world is doing well with GIMP, and the digital photographer has no choice but to leave Linux, or purchase Bibble. Not that Bibble isn't an excellent product, but it's never healthy to have NO competition.
I also believe that GIMP's implementation of color management could be the spark that wakes up companies that produce colorimeters, and perhaps kicks off system-wide color management through X, and support for ICC profiles in CUPS. I can only dream how cool it will be...
It's not safe to assume that clustering on its own will accomplish this. Most Enterprise cluster environments add significant administrative complexity. Given that human error is responsible for vastly more downtime than actual hardware failures this is significant.
The most effective means of improving availability is through automation of administrative tasks. For clustering to have a long lasting impact you first need a sufficiently mature IT practice. This alone can get you to a very significant level of "9's".
The second thing is adequate disk redundancy, and the third is environmentals. By this point you're at 3-4 9's consistantly and clustering is gravy.
I agree that the tools are (LTP) are rock-solid and critical to Linux' growth. I also agree that Linux is ready for prime-time, and aknowledge IBM's significant contributions to advancing Linux in the Enterprise. I even agree with LTP's purpose not being to proove anything.
My comment was specifically geared towards the article - not the tools, or the product of Linux. I think the article was very soft on details which is unfortunate given the talents IBM has working on Linux. The article seems to have its purpose rooted in adding credibility to Linux reliability, but I don't view the test as showing that. I believe you need a larger sample (test period) to make a conclusive argument about reliability, and a more wide range of tests.
I work as an availability engineer (studying reasons for downtime and finding ways around them) and the trends I see often take a year to show themselves. We won't even perform an analysis until we have 90 days of data, and that's rarely enough for a trend. An OS' reliability is based on much more than running a load of 90+% for X days. Very few systems actually run with that kind of sustained load in a production environment. Reliabiliyt is also heavily based on it's fault isolation, and resiliance to the unexptected (all of which Linux excells at IMHO).
I just don't think that this test did what it set out to do, although I'm sure marketing departments think otherwise. I would have covered various types of fault injection, and liklihood of human error for common recovery processes as well. I would also have had 90 days as my low end, not 30 days. Finally, I would have put these comparisons next to other options like Windows and a UNIX or two. I don't think the results would have changed, but I do think the results would be significantly more credible in prooving its capabilities.
I just finished reading the article, and I'm disappointed. It seems to me that the purpose of this article is to provide yet another "Enterprise-ready" stamp of approval for Linux. Now, let me be the first to say that I'm all for Linux in the Enterprise, but this article isn't the way to do it.
In the Enterprise world reliability equates to availability. This test used high loads over 30 and 60 days. To be Enterprise capable, in my mind, means 2-nines, or 99.0% availability. That's 3.65 days/year or 01:41:00 per week. To be impressive (and worthy of an article) I'd be looking for a minimum of 4 nines (0:52:30/year, 1 minute per week). Most mature sites I work with are closer to the higher 3-nines range on a consistant basis using a commercial *NIX.
What I'm getting at is that to have a real test give you a sample size that you can trust, 90 days is the bare minimum. I would have expected a 12 month study, and a much better explanation of Linux ability to deal with HW errors. IE - inject disk failures and scenarios which test fault compartmentalization. Doesn't matter what hardware you're running on - a good solar storm can play all kinds of havoc in a data center.
I'm disappointed that IBM would consider this anything more than a science project. Knowing the resources and talent they have at their disposal I was expecting more.
I'm wondering if the trend of having more workers telecommute recently is helping to keep the pricing faith alive for broadband providers. It's easuer for someone to agree to $40 or $50 a month if they're expensing it...
I would LOVE to find a way to convince myself that this just won't go anywhere for them. Unfortunately, I think it might work - take script kiddies as an example. They don't get the art of hacking, they just want to be able to brag about what they did (not how they did it). It's soul-less.
/) until I went to college and hung out with a dude who showed me slackware. Until you know there's another road, you never notice the bumps in the one you're on.
I think that's how a lot of computing is today. How it looks is very important to people who've never been exposed to the alternatives. I used to actually like Windows (I know - not good to say on
As much as Linux is making impressive progress, I don't think that road has been mapped yet for the masses. I hope I'm wrong though!
Maybe I'm missing something but it sounds like something that could be remedied by using page styles. Set up a page style with the numbering you want and assign it as needed.
I had different numbering formats I needed to use and just about pulled out my hair trying to find a Microsoft way to do it. Once I found a good book on SO/OO I fifugred out page styles and problem was solved.
Might be a different issue, but sure sounds the same.
Actually, there's more to it than that. I wish I had a better handle on the internals, but I have some documents in OO (SO actually, but same) which I can only PDF in the newer release. If I use an older release, or ps2pdf, the fonts come out jagged. Tried the same thing on Adobe's test drive on-line converter and they came out the same way - ugly. SO7, OO1.1 both convert nicely though. Very bizzarre.
No idea why some things work and others don't, but all PDF converters are not created equal. It would be interesting if someone who knows about this stuff could explain it.
Of course, M$ Office doesn't include Project Management software either... It's an EXPENSIVE add-on. Not sure how Mr. Project stands up, but RedHat includes it in its distro.
Reliability comes to mind, but I will concur that high levels of reliability in a workstation aren't needed by everyone. I have an Ultra 60 (2x450mhz/1gb) right next to an Athlon 2200 with 1.5gb. While I agree that the Wintel machine is faster, I disagree with the level of sensationalism you attached. I would also suggest that if you are getting 1mb / second throughput on a Sparc box you have some kind of a problem. Have you checked /var/adm/messages lately?
Here's another point to consider... Your 286 which you claim is fast as the Sparc box can't run Windows 2000. My Sparc 20 lab box is running Solaris 9 comfortably.
The list goes on, but your note is such an obvious troll I'll stop here.
SO versions 7 exported some ugly PDFs if you wewren't on a mainstream font, but the PDF engine in SO 7 is a completely new one with 3 levels of quality (Draft, Print, Press). I can't explain it, but when I demo'd Adobe's latest release it still produced jagged edges on one of my typefaces when rendered to PDF, but SO produced it cleanly.
Not sure if that's what you referred to by ugly, but if so, it's fixed. If you were talking about the interface in general then what the heck do you call attractive??
Is Sun really on the ropes, or is that what the analysts are saying because they don't "get" Sun's Linux strategy yet?
Take a long hard look at the R&D budget and the cash in the bank and I think you'll find that Sun is alive and well - even in comparison to most of its competitors.
This is flat out wrong. I use SO on my work machine and OO on my home machine. The only significant difference I encounter is in the standard fonts bundled. More information on the differences can be found in the OO FAQ.
You are correct in that the database component is only included in the StarOffice version, but is that really so bad? If it's such a problem than try this command: "rpm -ivh postgresql". It's not like Adabas is a MSAccess killer at this point anyway.
Let's put this in perspective: Sun bought StarDivision, and opensourced it immediately - that was pretty nice in my book. If you remember, SO 5.2 was entirely free until the business customers of Sun expressed doubt in the longetivity of the SO business model and requested that it be for a fee. Sun complied, and left OO in place as a free alternative for those of us less concerned about the business model. Oh, and SO was free for educational use.
I have trouble believing that someone who claims they are not very similar has spent much time working with them, but I suppose it is possible. It always bothers me to see FUD in the open source world; Don't let the dark-side cloud your judgement young Jedi.
I agree... I'm just not sure that the marketing organizations which are hungry for the "value add" would support that slim-down approach.
I also see where you are coming from in terms of calling software a product. With the right definitions in place your Honda analogy makes sense, but it would probably cause heart failure for most software businesses to think that way in our current litigation-crazy society.
The big point in your statement, which I think most companies wouldn't go along with, is that it costs less to design for reliability. Let me say that *I* agree with you. My supporting evidence for doubting business follows this line of thought is the emphasis on quick turnaround and time to market vs. quality. Short term gratification is what drives most businesses because they aim to please financial analysts more than consumers.
Look at the Ada programming language... It was designed to facilitate more reliable programming of large scale projects. Yert, most programmers don't like it because its "a pain in the a**". Yet, if you look at statistics, a cleanly compiled Ada program is much less likely to have bugs than a cleanly compiled C program. I know there's more to it than that, but my point is that getting engineers and businesses to look at the long term is going to be very difficult in today's hemorrhaging tech-sector.
I'm not holding my breath waiting for that to happen.
Couldn't agree more!
You make some good points here. I certainly don't have your experience - my only "stick time" was a college project on a RT-Vax, but I'm not entirely ready to abandon my thinking. My (revised) thought would be that most RTOS environments used for process control on things like laser vision correction and Boeing 747s were developed to much higher standards of both QA and architecture than an average PC operating system (like Windows). So, in some ways that prooves your point.
Because an RTOS will often have a human life at stake it has to bedesigned to higher standards. All that extra testing, design, and higher-skilled engineering time costs significantly more $$, and may be enough to make the costs of an OS prohibitive to a commodity PC. I would imagine that an extra $100 per PC license would have a big impact on how many Dells came loaded with Windows vs. Linux. The mere fact that an OS may have lives at stake creates a different programming atmosphere - one of responsibility than not all software engineers are up to.
Another point I would suggest is that an RTOS isn't doing as much as a regular OS. It would typically be stripped down to the essentials for its function. In contrast a modern OS is doing all kinds of things at once with combinations of software not anticipated by its designers, thus the integration is a crapshoot.
I think this lack of control is probably the biggest deterrant to OS manufacturers; If they were the only ones designing software for their OS they probably *could* warrant it. Unfortunately, who's to say what quality went into an arbitrary package loaded on their OS? You would never see a third party screen saver driver for a lasik machine (ok, bad example, but hopefully you know what I mean).
I do agree wholeheartedly that the average consumer has allowed themselves to tolerate much more sloppy coding than they should - we should hold vendors to higher standards (and hey, I work for one of those vendors!). The most obvious example of this is how many people tolerate needing to reboot their machine at least 1/day. Some people are amazed that my Linux workstation goes 100+ days without a reboot, which is sad. But I think it will take a few years of cultural inertia breaking before we could see an OS which could be warranted as a product. It would take a change to the mindset of engineers, engineering managers, marketing (who drives feature roadmaps) and executive management. That's no small feat, although definitely possible.
I think headcount makes a lot more sense than the more common per-CPU licensing we see on some enterprise software. Remember that the per desk fee includes Professional Svcs to make the transition, 60 hours/week of support, training. Also note that its per desk, so the gardeners mentioned in an ewarlier post wouldn't be counted ;)
Compare all of that priced out to what it costs to run oracle on a 20 CPU server, then imaging you have a DR site, a cluster with at least one more node... You get the picture. For a lot of sites this will be a huge reduction in cost.
It will be interesting to see if this experiment works. I agree with you - it may be viable for OSS as well.
Seems like there's a pretty big difference between something that is mechanically engineered and something that is software engineered. The term "product" IMHO applies to something that can be mechanically engineered (constraining myself to the Chrysler vs. software here). If you try to apply it to SW engineering I don't see how given today's technology it could be warranted as a product.
Chrysler doesn't have to depend on compilers and virtual machines and underlying hardware in the same way that software does...
I think you have a valid point lurking in there, but the analogy is lame.
I don't think I implied in any way that I'm a God - I only said that I do a lot of presos so that you'd know I wasn't a casual user. I certainly made no implication of your abilities, although your response certainly does. I'd like to hear more "facts"... Unfortunately your response doesn't include anything but conjecture.
You mention that "Corporate uses PowerPoint"... I wokr for a very large, very global company with all the same corporate red tape other large companies are hindered by. We use SO as our office platform and interperate with our MS Office using customer base with no significant problems. Many large goverments (talk about corporate red tape!) as well as educational institutions are also getting the message recently.
One of the things my team does is organize global customer forums where the customers present their best practices. Most of these presentations are in Power Point, and with few exceptions (mostly due to font translations) they load up in SO with no significant issues.
The presos which we deliver are authored in SO7, so obviously need no conversion. If you have specific areas which you can point out in which PP is superior to SO I would be interested in hearing about it. My observation is that very few of those areas amount to more than a a "cooler way to embed a MIDI file into my slide". I'm interested in learning about other areas as I haven't spent as much time with PP.
As for the spreadsheet, the only area I'm aware of SO falling significantly behind Excel is in 3d surface plotting. When we were graphing performance scenarios for enterprise storage we ended up using Excel for twiddling and GNUPlot for the automated suites because they could do it and SO couldn't (see, I can be objective on the negatives as well!)
Not sure where you picked up the negative vibe, but if you can carry a conversation I'm game to learn what you know. The question is, can you open your mind?
I was pretty happy with the functionality of Word 2.0 in terms of features. I'm not sure that pseudo-AI in the form of a talking paper clip has improved my writing skills.
:)
I think the pricing of something like an office suite (which IMHO is a commodity) is insulting. I think M$ would actually benefit more from including it free in Windows as a value-add than charging astronomical fees. It would be yet another notch in their office ubiquity and even further solidify it as a standard. What's more, it would bring Windows 1 step closer to competing with Linux (Linux does so much by virtue of included apps, where Windows out of the box does very little until you install apps).
Of course, I'm perfectly happy watching Linux keep building its lead, so Mr Gates can feel free to disregard my advice
I don't agree. "got to have powerpoint" makes no sense in itself.
For most of the presentations I do (and I do a *LOT*) I ignore all the bells and whistles and focus on my content. I put the bells and whistles into the way I convey the information rather than the file itself.
What I end up with a clean presentation file which exports very nicely to PDF where it smaller in size than the source file, and can be read on any machine, including those not powerful enough to efficiently run a complete office suite.
I've said it before, and will say it again. If Powerpoint is needed to make a point, then there are serious communication skills lacking in the presenter.
Us OO instead. You won't see much difference between the two.
The other point, which I think you intended but didn't include explicitly, is M$ format problems.
If Joe student bought a copy of Excel 98, and the book ships with a spreadsheet in Excel 2003 that student is probably screwed.
If the book ships with a copy of OO, they (as you suggest) can allow multiple platforms to access that file without worry about format compatibilities.
It's amazing how Apple has been slammed for obsoleting their customers, but M$ doesn't get the same bad press. I have a site I volunteer at which has to maintain licenses for Acess 95 and Access 2000 because they have Access 95 apps which don't work properly in Access 2k and vice-versa. So much for the benefit of backward compatibility.
This is one area which I find frustrtating. For all but the most advanced power users, OO/SO really provides what's needed for business functions.
The only exception is MS Access, however I think a lot of Access apps out there would be just as functional in a spreadsheet since the average user doesn't "get" relational theory, and doesn't program.
When the bastardized M$ XML appears I think we'll have an interesting potential for phenomenon. Historically, (as much as I hate to say it) M$ could do what they wanted without fear of repercussion because the alternatives weren't upo to snuff. Today a company has the very real ability to choose SO/OO as their office platform and take the initiative to ask their sources to use more open formats. This could be the first time customers of Mr. Gates can actually say "no" and choose an alternative.
Does this sound unreasonable? Sun Microsystems uses SO as its office platform (not surprising!). To be fair, I should admit that I'm an employee of Sun, but I think I can make fairly unbiased observations. In the 2 years that I've spent with SO,OO, and all the betas in between, I haven't yet encvountered an insurmountable document interchange issue. I have spent a lot of time writing technical documents, so I'd consider myself in the "power user" camp.
The most common "yeah, but" is presentation macros. My response: If you can't make your point without powerpoint animations & macros, then you don't have a good point. A good presentation *can* be communicated using PDF slides created by the M$ originator if it doesn't convert cleanly into Star Present. What if your customer doesn't have SO? Send 'em a nicely rendered PDF - SO7 has a completely revamped PDF engine which runs circles around SO 6. Better yet, use TEXT where its appropriate, and RTF in other places. If you're really hooked on format abuse, SO7 can even create a Macromedia Flash presentation.
People have come to make 2 mistakes because of features in common office products:
(1) relying on gimmics to communicate rather than relying on mastery of communication skills.
(2) assuming that just because everyone else saves in a proprietary, short lived format, that they need to also.
Given that so many other aspects of the tech industry are moving to value "good enough" technologies, I think we're very close to the tipping point for Office platforms as well.