The Death of Linux?

I got about halfway through Dirk's message and started to reply in a huff that I had never chosen the stinking Mac OS -- luckily I got all the way to the end before I sent my huffy missive. I almost just quit early in my LINC days because I was hating this stupid laptop so much. I've gotten used to it, but I am actually scheming to put Ubuntu on another laptop here and start trying that out again. The rapid fire releases of OSX versions have raised some doubts for me about how long I can keep up with the Mac OS without going broke. I think that to understand why Linux desktop applications are so damn ugly, you have to have a better handle than the author on why people develop open source software and on how much good, intuitive user interface design costs. There are a lot of reasons why people develop and contribute to open source software, but the most common one remains basic need. I spend a fair amount of time engaging with the Drupal and CivicSpace developers because I need that software. It shouldn't come as a great shock that there are a lot more competent software developers who really need a more robust webserver application than need an intuitive and user friendly spreadsheet application. As a result, Apache and MySQL and PHP are very solid packages while OpenOffice and the GIMP are still seriously frustrating at times. (For the record, I think Mac's Mail is one of the most limited mail applications I have ever tried to use. Mutt is easier to work with.) I don't think the problems with OpenOffice and the GIMP have anything to do with its developers Unix roots, I think the problem is that the OpenOffice business model doesn't allow them the kind of profits they would need to expect to be able to pour money into user interface R&D. I am really not that worried about the impact of this transition on Linux desktop use. I seriously doubt that the world is teeming with software developers who are trapped in Linux land now but will start developing for the more profitable Mac OS just as soon as Apple moves to an intel chip.

I had a lengthy reply to Dirks question planned, but Arif and Amanda have basically stated my thoughts. My short answer is that Linux will NOT be appreciably affected by Apple's move to Intel, primarily because OS X will still only be usable in Macs. Apple has stated categorically that they have no intention of licensing Macs to run on generic Intel computers. This means that using OS X still requires a commitment by nonprofits to purchase Macs. To me that's a no-brainer, if the nonprofit has the money to purchase and maintain Mac systems. Still, most nonprofits will always be short on cash, and if Linux remains a viable alternative on the desktop, your current strategy of propagating Linux use throughout the nonprofit/activist world remains sound. In fact, maybe it's the only choice, of necessity. I think Linux faces some self-inflicted dangers, but Apple's move has little to do with it. Will Linux continue to improve enough on the desktop to be an asset to groups? Or will its sheer geekiness continue to frustrate groups who lack easy access to tech support? Having used Linux, and as I follow continuing developments, I think that turnkey installations can be used successfully by anyone, even though the interface remains pure shite. But once something breaks, or new apps need to be installed, or a "simple" database needs to be created, then Linux becomes an unhappy choice for non-technical users. But then, poverty necessitates many unhappy choices. I share Arif's desire to just get stuff done. No environment has ever been better for me in that regard than OS X. But I'm not a professional tech. Your needs are different from mine. You may need a more robust mail program than Apple's Mail. Or you need a level of interoperability that I don't. That's understandable. But I don't share your needs. Mail just works. Everything just works. Still, I'm sorry in one sense to be using a Mac because I really wanted to be using Linux. Whatever Stallman may be like personally, I like his credo regarding sharing. I was under the impression that here was a movement that embodied many of the social values that I champion. Later, I was to learn that the majority of OSS programmers are actually employed (or otherwise contracted) by regular for-profit companies who are supporting OSS for strategic reasons (not least of which is to bring down Microsoft). Still, a great number of open source programmers share those values regardless of who they work for, and a wonderful thing has been given to the world. Sadly, all the Linux distros I tried were disappointing to this user. Now, I'm technically astute enough to be a Linux Guinea pig without freaking out, but my interests are strictly as a user (i.e. I really don't get off on digging into the bowels of a system and endlessly tweaking stuff). So from a user's perspective -- similar to the nonprofit desktop user you all support -- Linux really sucked big time. I understand the Linux developer mindset, and the origin and evolution of the Linux ecosystem. It's admirable. That's what made me want so much to use it. But you know what? That's not enough. I just want to get stuff done. As for the limited resource argument, I agree with Arif that with IBM, Sun, Novell, and others throwing serious money at Linux, the user experience should be a LOT better than it is. Every program thinks that it's idea of an interface is the best. And why not? There is no one playing cop. There are no unifying standards. No one creating a unified look and feel. Yes, I can spend a lot of time customizing this and that, different skins, yada, yada; but no matter how nice the desktop looks, Open Office still looks like crap. And don't even mention XMMS! Aesthetics matter, they are part of what makes a computer easy to use. I tip my hat to the Mozilla people, who have made Firefox and Thunderbird look good. Now if only the Linux community as a whole would hire their graphics people so there was a consistent look among all programs ... (I read where there are committees looking at this, but I've yet to see the results on the screen.) There's no polish, no elegance. It doesn't even look or act as good as Windows 3.1 let alone OS X! As for internal quality, you all know the drill: run the sound card setup from the CLI because it wasn't recognized by the setup program; hope the printer driver you chose works with your printer, because your printer isn't listed as such; peck through menus only the writer of the program can understand; three different and exclusive setup/preference programs; etc., bloody etc. I'm sure you know what I'm talking about. Amanda, I understand why you, a geek, would prefer Linux (although I don't know if its philosophical -- free versus proprietary -- a matter of personal taste, or if there's actually something important you can do in Linux that you can't do in OS X -- I'm curious). Of course, when we say Linux, we're not really talking of the kernel. We're using it as a shorthand for the GUIs, the apps -- all the stuff we actually interact with. It's a Tower of Babel. Short of a Steve Jobs dictator emerging on the scene -- not quite in keeping with the whole Linux gestalt -- I don't see how all these parts can be brought into a coordinated whole. Linux as it exists today is certainly usable, but with a high cost of time and frustration. Linux is built by tool makers for other tool makers. There's great pride in these toolmakers, and much of what they've done is excellent work taken separately. But it's not enough. Even Microsoft has shown better understanding of how to build a user interface. Anyway, there's no secrets. Both Apple and Microsoft have published interface guidelines for years. They're not perfect, but they're mostly good. The open source community doesn't have to reinvent anything. They just need to get serious about applying the collective wisdom of a whole generation of great interface designers. Otherwise Linux is just a circle jerk for programmers, and the rest of us are left to fend as best we can. Let me end by addressing two statement of Dvorak's that I think are completely stupid and have led to confusion. "While Apple ran on the PowerPC chip the amount of developer effort in the Open Source camps was nil. But now that Apple is using the same processor as everyone else, targeting the Macs will now be an easy decision to make. This will be at the expense of Linux." - Dvorak The processor is largely irrelevant (especially if you resist the temptation to get a speed boost by inserting chunks of assembly code into your app). You still have to program for the operating system. It will be no easier to port a program to the Mac than it ever was. This change to Intel will prove absolutely no difference to anyone except to existing Mac developers who may have to re-write their apps somewhat depending on whether their program is Cocoa or some flavor of Carbon. There is no technical reason why this will suddenly make programming for the Mac easier for the open source community. As an example, consider OpenOffice. OpenOffice really sucks on the Mac. (I'm not saying OpenOffice as such sucks, though I'm not a big fan personally. It's just that OpenOffice running under X11 has nothing to do with the Mac experience. You might was well be using a generic Intel box running Windows.) The OpenOffice developers have abandoned the Mac port because in order to have OpenOffice run as a native OS X application, it would take a significant effort for the OpenOffice developers to rewrite tons of stuff to address the Aqua layer and other parts of OS X. This has nothing to do with the CPU and won't be alleviated one bit by a switch to Intel processors. (Long story, but you can read bits and pieces on the net). Fortunately, for those who care, the NeoOffice project, which was working simultaneously but separately from the OpenOffice effort (they use Java to interface with the Mac GUI bits), has done a good job of making it suck less, and they plan on doing more work to get 1.1 as fully integrated into the Aqua GUI as possible. And since they're using Java, it should be a cinch to make their efforts work on the Intel version as well. A NeoOffice edition of version 2.0, however, is a long, LONG, ways away (if they ever do it -- it's mostly just a couple of guys doing it out of love). Anyway, none of these challenges have anything to do with the underlying processor. There are 30 million Mac users and, what, hundreds of millions of Windows users? Only an Adobe or a Microsoft have the resources to run parallel development efforts (let alone are they going to duplicate their efforts again for Linux/Gnome, Linux/KDE, Linux/??? -- this gets back to my rant about a single unified interface). "Linux was the only X86 alternative to Microsoft and now it has both Bill Gates and Steve Jobs to contend with..." - Dvorak If by this statement he means that the Intel box is a blank slate, on which a user can choose to run Windows, Linux or OS X, then he's an idiot (but we know that already). Linux STILL has only Microsoft to contend with. It remains a contest between Windows and Linux. And the Mac remains a proprietary product, hardware and software, that JUST HAPPEN TO USE an Intel processor. Nothing more. The Mac lives in a different ecosystem. Those users who want to load an alternative to Windows on their existing Intel computers will still only have Linux as the main choice. If you want OS X -- more generally, if you want the Mac experience, with all that implies -- you have to buy a Mac. Apple makes their money selling hardware. I don't know what else to say that would be helpful. Good luck. It's not easy.

I am a stranger in these parts and have only dropped by to read mail regarding Linux on laptops. It is a little difficult to judge when the last of your missives were written because criticism of a lack of support for Linux desktops could date your dispatches from 2000-2005. I began with RH 5.x and currently beat my brains out with Suse 9.x, Redhat 9.x, Fedora x.x and Mandrake 10.x. My degree is in electrical engineering and my first computer course was in the 1960's for true magnetic core memory. My fascination with Linux has to do with my fascination with stripping my Columbia Bicycle down to to bare parts when I was 12 years old, tearing radios apart, re-jetting the carbueturs on my motorcycles, and all other things difficult and not necessary. Linux of course fits that paradigm perfectly. As for chosing it to bash Microsoft, I maintained from the beginning that Linux did not "blue screen" ever and that made it better than anything MS had. That is until Windows XP which I told friends was as stable as Linux. I considered that high praise for XP until Service Pack 2 for XP arrived and made XP useless for running older engineering programs. So now I am back to using Linux on my laptop because with WINE it is possible to run older programs which no longer are allowed to run in an XP SP 2 environment. My one remaining Linux problem has to do with getting a PCMCIA wireless adapter to work in Linux. I have both the Netgear MA401 and also a Lucent Silver wireless card. Neither of these work in a SuSE environment, which grieves me greatly since I splurged almost $100 on the professional edition of this product. Weeks have gone by and I am vexed to understand how a wireless card can be assigned IP's properly but still won't allow access to the Internet? Since these cards worked with XP I know it is neither myself, my wireless AP or my DSL provider. Tonight I will be removing SuSE from the laptop and going back to RedHat 9.0 because I know it does work with these cards even though I still don't know why? Perhaps the above adds to the handicaping percentages regarding bets being placed as to whether or not Linux can survive for home users? I love Linux as passionately as I have loved women in my past who have burned me to a frazzle too. I am older now and married to a good woman who does not frustrate me with undecipherable challenges. Whether I can say as much about my relationship with Linux in the future remains to be seen? Please excuse me if I have butted into something more private than I imagined. My addition adds to the noise but not to the data thus I probably should have not sent this. Thanks for reading it anyways. John Casapiedra, IEEE 40270636