The Linux Desktop Obituary
rcriii writes: "Kevin Reichard is announcing the end of Linux on the desktop over at Linux Planet . Having spent the past couple of weeks fighting with Star Office and Netscape, I'd say that he has a point. Let the flame wars begin." I'm still not sure it was ever born in the first place ... although I happily run Linux on all of my desktops. But I'm not exactly the desktop of corporate America either.
For someone that runs Linux on all their desktops, you certainly seem to be playing a lot of Windows only games every time a story gets posted.
Many people have commented that the arrival of Mac OS X will make it easier to to port Mac apps to any Unix, since "developers will be forced to port [apps] to Unix". This is totally inaccurate. Mac OS X doesn't have anything approaching a Unix API (although it is Posix-compliant, that only consists of the BSD layer. All the fun stuff that makes the computer usable, like the GUI, is Apple-prorietary). One of the main APIs, Carbon, is a slightly altered version of the traditional Mac OS APIs. The other, Cocoa, is basically NeXTStep's OOP frameworks. A program from either of them is just as hard to port to Unix as before; cocoa programs are probably harder.
..install their own systems? Outside of the IT departments, almost no one does anything on their machines other than point, click, and print.
Don't try to use Linux as if it were Windows. Windows does that better. The fact that most companies have (or should have) a pervasive network by now, coupled with the fact that each end-user workstation probably needs to perform at most one or two tasks, makes Linux perfect.
One appropriately skilled Linux hacker could probably run an entire small/medium-sized company (about 100 employees) if you left all of the software/hardware decisions up to her. And I don't mean by having her install crap like StarOffice on 100 Red Hat boxes. I mean by asking her to analyze your business and having them suggest the way to make it all come together, with modern technology. If you step back far enough, every system is an embedded system.
The fact that you can totally control Linux means that you can design a more efficient end user experience. If the Linux hacker is appropriately skilled, you needn't even employ them fulltime, ie -- they're a consultant!
This is what Linux companies should (and do) focus on. It is total silliness to try to provide something that looks and acts like Windows, when realistically, Windows is totally ill-suited to a lot of these tasks. Copying Windows means you're just doomed to fail. They do it better, and they have a monopoly. Instead, do something innovative
Windows has it's places, but Linux has so many more.
The following isn't intended as flamebait, although my tone might suggest it ;)... Having killer apps under Linux isn't going to make it anymore appealing as a desktop. I don't know whether it's the flexibility of X or just plain incompetence by the applications programmers for X, but the UIs (including when running KDE) really suck. It's as if the applications programmers are completely clueless about these things. Don't knock Windows (or Mac OS) for how generic or inflexible their UI is... it works, it's consistent and it's far less quirky. It's the quirkiness that really causes problems: for every application you have to learn how it behaves, and how it differs from others (even though it looks the same). And for crying out loud... test the tabbing order in dialogs!
<current pet peeve with X applications>
MAKE YOUR MODAL DIALOGS COME UP CENTRED OVER THE CORRECT APPLICATION WINDOW! I was playing around with Mandrake recently. I had my application window (Netscape or Mozilla, or something) in the lower left of the screen... every dialog appeared in the top-right (which is a long way away at hi-res). In fact, the dialog placement seemed almost random. What with the other Windows on the screen, it's often hard to notice a dialog pop-up when it's so far away from the action.
</current pet peeve with X applications>
Linux advocates always seem to get caught up in this "What one app do we need to make this popular?"
The thing is, it isn't one app. It's 40,000 applications that make Microsoft Windows popular.
You have to have an environment where anything that is needed, especially anything sent to you by partners or regulatory agencies... just runs.
I used to run into this same issue back in the days supporting OS/2. At the time it meant we had many users with two desktops. One to run OS/2 and our "standard" apps, one to run all the other apps that didn't work under OS/2.
This is expensive and a nightmare to maintain.
Anyway, it's not just one app... It's everything.
My Linux desktop has never been this usable, and it keeps improving by the day/week/month. The good folks at KDE have been doing wonderful work! And the XFree86 gurus have really picked up the ball the last couple of months. From my own experience I;d say we're only starting to compete. The fight is not over... and I'm confident Linux will be one of the last ones standing at the end of the battle.
:)
:)
Concluding that the Linux desktop is a failure simply because a company that writes a silly Linux/Unix file manager for it fails is really, really stupid! Eazel's business plan was flawed beyond believe! They counted on revenues from "services". However, the services they provided were already commodity items (supposed proper package management, free internet disk space).
And then there's this quote:
Such is the way of all movements: either the professionals take over and the movement evolves, or the movement recedes.
No shit! So the Linux desktop attracts some cheap money, but when that money fails, so does the Linux desktop? Give me a f*cking break
The only thing that's dead or dying are companies with flawed business plans. And for the record, I don't hope the author of this articles tries to make a living by writing, he'll starve
-adnans
"In short: just say NO TO DRUGS, and maybe you won't end up like the Hurd people." --Linus Torvalds
I'm an RHCE. I was down at SSC (the Linux Journal guys) filling out some paperwork, and I needed to pull an IRS form off the web (I had forgotten my copy). So I'm telling the secretary the website, and we get the form, and Netscape cranks up Adobe Acrobat Reaer, and now we go to print, and... oh, RIGHT, of COURSE it's a Linux box, this is SSC!
I had made the mistake of thinking "secretary == Windows box" and had not twigged to the "K" on her toolbar until the print box said "lpr".
Now, remember, I've been running Linux since 1995, I'm supposed to know what an X desktop looks like. I couldn't tell the difference until it hit me in the face.
Linux not ready for the desktop. Feh. My wife has run Linux on her desktop for three years now. Word Perfect, Netscape, Solitare, Free Cell, Minesweeper, now Hearts is out, Quake, Civ, Gnucash, GIMP, and Samba if you must talk to the EE... she demanded Linux after Windows 95 ate the registry twice in as many weeks. She hasn't missed a thing.
The ONLY thing Linux needs is shelf space at Comp*USA for preinstalled machines. Other than that, we are SO ready for the desktop. Anyone who says otherwise is a liar or a wuss. I count Microsoft marketroids in the former category.
--
Software is like sex.
It's better when it's free.
-- Linus
It is a fact that Linux faces many obsticles on the desktop. Although it is catching up fast, there is still a gap in productivity and end user applications, it is still harder to setup and use for non-technical users, and it still lacks certain functional consistency (look at font handling, for example, and I'm not talking about anti-aliasing, but consistency across screen, printing, and applications).
I am 43 years old and work for a company run by 50 year olds. I'm one of only a couple of Linux users around here, and certainly the oldest one. It will be a long time before I can convince the company management to switch desktops to Linux, expecially since there are still perceived and real shortcomings.
But Linux is very popular among technical college and university students. What will happen in another few years when tens of thousands of college students who have grown up on Linux go into the workforce. Certainly some of them will either start their own companies or move into decision making positions in existing ones. You will begin to see Linux on the desktop, and increased demand for Linux applications. The timing will be very good for this because by that time, most of Linux's desktop shortcomings will have been addressed.
Success (if not dominance) on the desktop will take the longest of any area, but it will eventually happen.
Linux on the desktop is great for those of us who use the full functionality it has to offer, or at least a subset of it.
Your average office user though WANTS Windows. Okay, it may crash quite frequently, but let's face it; it's intuitive, and well-designed aesthetically.
Lets keep working on the desktop for ourselves and if others want to use it, great... But lets not forget where our true strength lies: In the Server market.
Bzzzzzt..."AAAAaaaaarrrgh!!!" Thud.
Let's see:
Linux on the desktop is dead because the desktop applications, in development for 3 years or so, are incomplete and immature.
Linux on the server is alive because the server applications, in development for 6+ years or so, are complete and mature.
So, linux on the desktop is dead. As in, incapabable of life. As in, permanently deceased.
In a couple of years, these applications will still be incomplete and immature... um... why?
Personally, I think Linux on the desktop is in it's infancy, rather than it's deathbed.
Microsoft Windows always has been, and always will be holed up in the desktop market. It's stupid for Linux to even think of competing there.
That's the statement of a troll, I think, but I'll pretend otherwise. 5 years ago, I had to compile everything for linux. I didn't even try to get X running. There was no office software, at all, that I knew of. Mail was text only. Fast forward to today: several office packages, 2 of which are free, to choose from. I enjoy having mutt as a mailreader, so I read from a shell, but it has both inline pgp and can spawn up attachments into StarOffice right from the mail client. Oh, and I don't have to worry that every attachment I'm getting (or every email, for that matter) is some VBScript laden bomb waiting to infect my system with a virus and wipe out all my files.
When I need to set up a new desktop client, it typically takes 3-4 days, using Windows 2000.
About 6 months ago, I needed to update my Linux box (well, wanted to, anyhow, but that's such a fine line for a computer geek
An inferior operating system, matched by shoddy programs that don't work
I've spent months now using netscape and staroffice every single day. Netscape is utterly stable, when you discount Java, which I have off on ALL platforms due to that instability. Try comparing the resources used by the latest linux netscape to those used by IE. Mozilla isn't a 1.0 version yet, but I tried out 0.9, and it seems like its well on its way to being nice.
That said, let's look at what I get that you don't:
I get built-in firewalling/packet-filtering capability, and RedHat 7.1 will even build access lists for me.
I get a variety of mail clients, graphical or shell, which are faster and more robust than garbage like exchange (ever try to close a multi-thousand message box? Better have something else to do for a while), as well as not being easy victims for every virus writer on the planet.
I get system monitors, media players, development tools, games, graphics programs, irc clients, a palm-sync package, cd ripping/burning software, newsreaders, as well as a desktop that is so customizable it makes the pathetic attempts under windows ("Oh, your mouse pointer is now really a mouse, how cute *gag*") worthy of tears. And of course, I get to choose from a variety of office suites that are free, never mind the commercial ones.
Meanwhile, Linux as a server has grown from fun to just amazing. With a vastly improved kernel, now offering fast context switching, and speedy multithreaded I/O, you get software like Apache, PHP, MySql/Postgres/Oracle, sendmail, BIND, etc, all of which put their Win2k equivalents utterly to shame (with the possible exception that MS SQL Server is pretty decent). Want an incredibly fast webserver? TUX in the kernel and you're beating the pants off anything.
This is to say nothing of the true cost of ownership. Imagine an IT world based on linux: you can export X displays without expensive add-ons, you can ssh safely and securely into your client machines, you get built-in packet filtering with logs you can easily transfer and audit, which your users can't override at their whim, AND is fast and free. Linux really COULD be a desktop now. I shudder to think how amazing it will be with some market share. Add to this the incredible stability -- I'd like to see you put a windows box at a colo provider 1500 miles away and feel safe when you walked away!
CompUSA cheap... nice troll, or nice setup, but give me a break -- IT people spend 10x as much time just repairing virus damage on windows garbage than they would converting their whole company to Linux. Games under linux are coming, and when they do, Windows is finished.
Unfortunately due to the /. effect I can't read the article.. But I've always had discussions with coworkers on this topic.. Linux has yet to be a user-friendly desktop. And it probably won't be for at least another year. Sure its the desktop of choice for geeks and techheads alike, but imagine installing a fresh copy of Debian on a Pentium, and giving it to your grandmother? Unless you've done a really good job of making sure everything is working, and you show her how to use things, where is she going to start?
.. and don't get me started in printing in Linux.. sigh.
Obviously she can't call the guys at GNOME for customer support, or the guys at KDE to ask why Konqueror isn't rending a webpage properly. It these things that are preventing it from being a true workstation for the masses.
Not only that, Linux tries to mix Server and Workstation too much.. Once again, the average geek will like this, but most people don't care if they have a telnet server running, in fact its a huge security risk for the average home user.. Considering he'll probably be storing webpage passwords on his machine.. Then there is lack of a good Web Browser, although this will soon be a thing of the past, as I've been using Konqueror myself for months without a problem.. But Netscape bundled by default is horrible.. And then the one topic that is keeping it from being on every machine, is games.. Loki is doing a good job trying to fix that, but even I had trouble getting Quake 3 running properly with a PII 450 and a Voodoo3, It was slow as hell, despite talking with reps at Loki on which Mesa libraries to use and install, only to get a "Well we don't really know" answer.
Then the common answer for people that are struggling with Linux and always asking questions is "RTFM", well guess what, there are people out there that don't want to learn about a computer, but just use it. And futher more, I doubt this person has a book on GNOME, and people trying to learn GNOME aren't going to know GNOME has built in documentation, or what the f*ck is a manpage. I can say the same thing, I don't care about how my Microwave or Toaster works internally, but when I put in leftovers or bread, I expect them to be heated and toasted..
Thats the only thing I give Windows, I can install it for my parents, show them the icon for IE, put a few games on for my Dad, show him the icons, show my mom the "Word" icon, and how to print, and they're set, happy and have little problems.. I only need to teach them when blue screens pop up, or things lock up, press the reset button and start over.
..There's a-dooin's a-transpirin'
--I have Linux on almost all of my machines. I have one purely WIN box in my house and my wife uses that.
.zip files and have them do their job.
--It started with, "I don't care if it's better, I can't run Quicken. I don't want to run something as good as Quicken I want to run Quicken. I can barely see the words in Netscape. Why can't I play The Sims on it? Nothing works. I want Office, I hate Star Office it is ugly. Dell says they can't help me becuase the machine came with Windows 98."
--I gave up after that. She's not a stupid user either. She's a power user for the NT set. She just wants things to work as expected. tar -xvf doesn't make her happy. She like to click on
--I love Linux and tweaking and such but guess what I can't play NASCAR 4 on it.
--Linux never stood much of a chance
---
This
Just my thoughts, but...
Okay. Start with this: why the hell does kwrite automatically copy selections into the clipboard whether I want it to or not? (Did that on RH6.0, anyway; I haven't really paid attention to whether it's been fixed since).
I agree with those who say the Linux community is too insular. We sit here and bitch and whine that the world isn't taking us quite as seriously as we'd like, but who's out there hacking OpenOffice to help make a workable MSOffice replacement? Why are we letting politics dominate our desktop decisions? And why the hell isn't the Linux community trying to forge alliances with the Mac community?
I'll start with the last point first and get the flamebaiting out of the way. When the Mac came out, Apple put out the Mac Human Interface guidelines. Microsoft has its own rules for Windows. We have no such thing for either of the significant Linux desktops. Believe it or not, this is a bad thing.
For this to work we need some interface guidelines, preferably written by someone using both MacOS and Linux (since Mac users as a general rule are more sensitive to clumsy interface design) (ducking flames, please hold). Specs like this are not amenable to committee design, so they should be handed down from on high by the toolkit developers for public comment rather than designed by committee. That's one.
The politics have got to go. Yeah, you've got your favorite features; I personally rather enjoy the look (though not the feel) of Athena widgets. Stuff 'em; you can add them later if you need them, but you need something to fit them in with. We have two desktops, which is one too many, and they will probably never be merged. Fine. Let's go with what we have and relegate the rest to the special-purpose bins where they belong; Motif is dying anyway, and it's the only other toolset that really counts. And if we must keep them, let's have XawGTK and QtLesstif around so we don't get confused and have to look at fruit-salad apps.
As for the matter of open source desktop apps, we only have ourselves to blame. Browser? Okay, Galeon's halfway there, but you still need Mozilla. That's ridiculous; AOLTW/Netscape changed the licensing, so there's no reason for separate downloads. Desktop? Sorry, I have no sympathy when OpenOffice goes wanting for developers. You've got a rather useful package there -- huge, but it's got everything you need and an open file format to boot. If you have nothing at all to contribute to OpenOffice or any of its competitors, you have no right whatsoever to bitch about articles like this.
The Linux world needs to swallow its pride and accept that some decisions do need to be made from above, or at least proposed from above and accepted by a critical mass. You fork, you're out. You've just created a new community, and the burden is on you to get it accepted, not to whine about why it isn't.
There is another component to this. Our desktop flagship programs are huge. Mozilla is about a 20MB source tarball IIRC, and I believe OpenOffice is well over 300MB. This is IMHO unacceptable in a Unix-based community; monolithic office suites are a Bad Thing to begin with, and given that there hasn't been a really core-type feature invented since the multidimensional spreadsheet I have to wonder where all this bloat is coming from (since I don't use it I could be off-base). Same with wasting space on skinnable browsers when performance should be the big issue.
We need more than developers in the Open Source community, you see. What's missing from the Open Source equation is support personnel like tech writers and creative people. We need more books like Coriolis Press' Lions-style source commentaries. We need interface designers willing to make stuff look pretty (something I'd love to help with, if anyone wants a Mac user's view, btw). We need Open Source RAD tools like VB or MacOS X Interface builder. It is very much time we reached out to the rest of the world to see what there is to be offered.
Yes, there's the marketing problem as well. Don't expect Linux desktops everywhere next year. But how many people who know only the hype are aware that Linux is coming up on its tenth anniversary?
/Brian
Was KDE. Eazel may be dead, but KDE continues to get better and better. Konqueror is arguably the best browser on any platform (and is at least as good as MSIE), and yet is still a relativly new project. KOffice is coming along in leaps and bounds, and given that the KDE team were able to make a Mozilla-beater in far less time, *from scratch* (Mozilla is based on pre-existing NS code remember), I have high hopes. It's already extremely useable for day-to-day tasks, and above all is quick.
I continue to use AbiWord for its MS Word importing features.
Linux isn't dead on the desktop, you just need to look in the right places.
--
Blaming GW Bush for the Iraq war is like blaming Ronald McDonald for the poor quality of food.
Been there, done that over on linuxtoday and linuxplanet.
- 05 -22-006-20-OP-DT
/ 33 92/1/
You will get this on slashdot tomorrow, so I am might as well post it now. Brian Proffit's rebuttal is here:
http://linuxtoday.com/news_story.php3?ltsn=2001
Original article is here:
http://www.linuxplanet.com/linuxplanet/opinions
Sorry, could not be bothered with html.
I believe Kevin did this about the desktop to rile people up and get them motivated. Whcih seems to be working...
StarTux
3-4 days to set up a windows 2000 desktop at $20/hr?!?!?! God, they need to get rid of you and hire me. At my old company, I could take my time and have a win2000 machine up in 4 hrs, even given only adequate hardware. That includes office 2000, service pack 2 and setting up their mail etc.
I got laid off last month, and here I am stuck at the fscking university making beans for pay. Seriously, your company needs to HIRE ME!!!