Kasperian Moving Parts

kinda like Batman, but with a wife and 3 kids

Having spent a few days with his MacBook Pro…

| 40 Comments

I recently blogged about Desktop Linux possibly having some core/fundamental problems that might be keeping it from enjoying mainstream adoption and 3rd party developer attention as compared to, say, OS X. To my immediate defense, I’ll say that it was actually more of a brain dump and rant (True Story!) than a well-thought-out dissertation on all of the issues at hand. The impetus in this case was:

  1. Frustration with a particular admittedly proprietary application that didn’t use to have any problems in Ubuntu 8.10, and since then has been nothing but trouble for me and roughly 90,000 other people. You may say that it’s unfair to fly off the handle at one proprietary application having problems and condemn all of Desktop Linux, but I do not think this is limited to only one proprietary application.
  2. A shiny new MacBook Pro in my possession and an epiphany of “this is what we’ve been working for, guys and we’ve been doing it for more than a decade and we’re still not there yet, why?”  And I’m not talking about the pretty UI or shiny buttons. You can argue all you want about OS X being the best in the UI/shiny/usability categories. That’s not what I’m talking about. I’m talking about the increasingly growing market share of OS X and the (generally speaking) more polished and well-thought-out and 3rd-party-developed/supported applications. Being able to go to Flickr, for example, and download an actual client for OS X is pretty darned cool. Sure would be nice if Linux had the same mind/market share.

Now, having spent a week with my MacBook Pro in both Linux and OS X, I have a few more thoughts to add to the fire. Some of these have been results of discussions had as a result from my earlier post on this subject and others are more related to time spent with said shiny new laptop. But I think these are more constructive and less inflammatory. =:)

  1. The MacBook Pro is truly a nice laptop and Linux, for the most part, runs really well on it. The Karmic wiki page got me 95% of the way there. Unfortunately, rEFIT doesn’t understand GRUB2 at all, it seems, so to just get Kubuntu Karmic to boot, I installed the old GRUB 0.97 instead (sudo apt-get install grub). Getting sound to work requires alsa-driver-snapshot, so that’s a little painful but not too bad. More painful was the hour I spent yesterday trying to figure out why sound stopped working (and this time it wasn’t pulseaudio’s fault, but rather something weird with the ALSA driver that plugging headphones in and removing them seemed to fix). Getting click+drag to work on the MacBook Pro 5,5 (since there’s no physical buttons anymore, but just one big touchpad (WHICH IS REALLY NICE!!)) requires a custom bcm5974-dkms driver. I’m using an unjournaled hfsplus partition to share data between OS X and Linux. And “setxkbmap -option altwin:swap_lalt_lwin” (or setting the same checkbox in KDE4’s System Settings) lets me use the command/apple/squiggly key next to the space bar as my Alt key (for alt+tabbing, etc.). All in all, I’m REALLY happy with Linux on this MacBook Pro. It seems to work every bit as nicely as Linux does on my work Thinkpad T61.
  2. There is something core to my nature that must tweak and hack, and Linux is most conducive to that. Take something simple, like wanting to change the font and font size that OS X uses for window titles, system menus, etc. Apparently you just can’t do it? That kind of stuff bothers me (and this is just one example in OS X that comes to mind). I truly do love Desktop Linux, and especially KDE for this reason. I’m not saying I couldn’t survive in OS X, and I still enjoy it and its apps. And if I have problems with things I need to do (audio/video conferencing comes immediately to mind), I have no hesitation booting into OS X to just get things done. And no, sorry, I just can’t stomach the thought of using Windows because I have to get things done. OS X may not be free, but at least I don’t vomit from just the thought of using it. But if for nothing other than the challenge of trying to figure out how to get things working to my liking, I feel compelled to run Linux on this little wee beastie. And maybe after I get things working to my liking, I’ll even find a couple of itches to scratch again and start being productive again. =:)
  3. We’re (Desktop Linux) not there (3rd party developer interest, compared to OS X and Windows) yet, but I think we’re getting closer, and even so, we may just never get there and that’s not our fault, I don’t think. My original line of thinking was that we’re not there yet because we keep changing core system components that prevent 3rd party developers, etc, from taking our platform seriously. And I think that as much as possible, we should really try to stop changing/breaking stuff so that this is not the reason we don’t get there. However, we have other core values and tendencies in Desktop Linux that are most definitely contributing to us not getting there and some of them we cannot change. Let’s take an important one: Freedom. Both parts of freedom matter: both gratis (for zero price) and libre (free to do whatever I want to it). The first part means that 3rd party developers can expect to sell nowhere near as much of their software in Linux as they can on OS X or Windows. The second part means that 3rd party developers can expect to meet resistance to their very existence. I personally side with the former more than the latter on this, since I’m cheap by nature and like to not spend money whenever possible. Also, being that I have to work for a living and support myself and my family, I do not, for a second, fault companies for existing and needing to have me pay for things so they can exist. But I think I’m in the minority on this point in Desktop Linux. Lastly, we’re not a money-making machine like Apple and Microsoft, and we never will be. And that’s both good and bad. It’s bad in that we do not have a big budget to spend on advertising and cute commercials, etc. It’s good in that we’re not going to go out of business just because we’re not “profitable” or growing as fast as OS X in market share. We have and can and will outlast other OS’s and desktop environments that must be profitable to exist (OS/2, Amiga, BeOS, etc., etc.). And maybe that’s why we’ll finally succeed in continuing to gain market share. Or maybe we’ll get there by being stable and good enough for most users and having applications which live on the  Internet being more important than applications that are written for Linux (Google OS, perhaps?). And maybe we’ll still not get there for another decade. Or longer. But that’s not the point, really, is it? I mean, it would be REALLY nice to never hear “oh, we’re just not even going to bother doing XXX on Linux, but that’s okay because Linux doesn’t matter… heck, it’s only .05% of our sales anyway!” again. But we’re here because we like what we have and we like where we’re going and we like controlling our destiny. I totally get that. And maybe that’s good enough.
  4. OS X really does have some nice apps that Linux has no counterparts for, but for me, there are not that many (Tweetie comes to mind) and the good news is that we (Desktop Linux) can fix that ourselves (and we are). One of the big things that people point to in this whole OS discussion is that OS X has more polished apps than Linux. And I must agree with this. But I think the reason is less because of OS superiority and more the nature of apps on OS X. Developers actually make money selling software on OS X *shock*, so they have a vested interest in polish and user experience, and spend a lot more time on it than Linux projects typically do. Heck, they even pay people to help make their apps polished and highly usable. The good news is that we’re in control of our own destiny here and can do (and are doing) better.
  5. We’ve come a long way (baybee), and I think we’re on the right track. It’s pretty amazing to think how far we’ve come in the last decade. It seems only yesterday that Blackbox and Bbkeys were the coolest thing in the world to me, and that the big Linux Desktop Environment projects just were nowhere as fun, exciting, or good-looking. We’ve come a long way since then. I’ve always preferred KDE to GNOME, but both DE’s have made HUGE improvements–both to the core Desktop Linux technologies that we share and to UI and polish and usability on top of those technologies. KDE4 has not even been out for 2 years already and the difference between what we have now compared to what we had 2 years ago is phenomenal. I absolutely agree with the sentiment that Desktop Linux is now, more than ever, ready for the world to use and ready for 3rd party developers to start writing for. I think this should drive us to be even that much more cautious as we push out new distributions and make sure we don’t break stuff just because we want something new and shiny.
  6. I haven’t left, nor do I plan to, Desktop Linux (I know, who cares), but I think it’s important to think about what might be broken in our development model and figure out how to fix it, else we’re shooting ourselves in the collective foot and preventing all our hard work from reaching beyond our little geekly communities into more mainstream adoption, and that would be truly sad.
  7. This MacBook Pro still has problems with Linux and that sucks, but also not our fault… sorta. And I’ll figure out how to work around all of them and so can you. As an example, as I’ve been typing this blog post, I have brushed the touchpad about 10 million times accidentally and ended up clicking somewhere entirely else, and in general disrupting my work by having to keep undoing the garbage that just happened. And yes, I’ve added a HAL fdi file to enable SHMConfig and tweaked syndaemon and synclient and yes, I know I need to still do more. But here’s the point: this does not happen in OS X. It just works. It would sure be swell if we could figure out how to make this work out of the box for Linux users. Similarly, suspend seems to work just fine, but I’ve had several issues with X/keyboard/mouse stability upon resuming. Not surprising, being that Apple hasn’t tried to make sure that its hardware works well with Linux. But annoying all the same.

Anyway, life is good, and I have a new puzzle to figure out (this MacBook Pro). =:)

Author: Jason 'vanRijn' Kasper

My name is Jason 'vanRijn' Kasper. I am the ring leader of the amazing Kasper family. I am unashamedly a Christian Nerd. These are our stories....

40 Comments


  1. We’re (Desktop Linux) not there (3rd party developer interest, compared to OS X and Windows) yet, but I think we’re getting closer, and even so, we may just never get there and that’s not our fault,

    3rd party developer interest isn’t there because Linux isn’t an appealing platform to them and it goes well beyond marketshare. There’s too many distros, windowing systems, sound apis, and worst of all the distros are designed to only distribute open source software. To build something like an msi for Linux takes far more work than it is worth, especially when you have to support not only multiple distros but multiple versions of those distros as well.

    Linux at its very core is anti-proprietary. The distros are a further extension of that. Until the people behind Linux acknowledge the need for proprietary software expect Linux to stay at 1% for another decade.

  2. Yep, good points, definitely. I forgot about the whole issue around software distribution. Absolutely spot on. That’s another nicety that I forgot about OS X… being able to download and install software that is guaranteed to work. Of course, that’s easy when you only have a single distro (OS X) to worry about. But still, good points. =:)

  3. Pingback: Mac Laptop News » Blog Archive » News: “Jason Kasper (vanRijn): Having spent a few days with his MacBook Pro…”

  4. @ Eric Raymond
    microsoft has windows xp, vista and 7 in the wild. Windows has win32 and .net API to develop against .. on the graphic front, you have directX and OpenGL .. microsoft break stuff too btw releases or all apps that worked on ewixp would have worked in vista .. how many windows devlopers have quite because of the mess that exist on windows?

    hulu just released one binary for all linux distros …one is packaged in .rpm and the other .deb but they are the same

    firefox release only one binary for all distros

    virtualbox release only one binary for all distros ..all specific distribution packages deals with extra stuff like where to put the icon on the menu list

    vmware only release one binary package for linux

    nvidia and ati release only one package for all linux distro

    the concept that “too many distros, toolkits, api” turn off 3rd party developers is simply not supported.

    developers dont target linux simply because of the market share ..we could give them cookies and they would still not bother if linux doesnt get market share or if we dont complain loud enough …


  5. microsoft has windows xp, vista and 7 in the wild

    Yes and there is binary compatibility between them.

    You mentioned some proprietary developers that release single files for Linux. What you didn’t mention is that they contain scripts to make adjustments for each distro. The distros lack basic standards between them, this is why the LSB was created (which most distros don’t follow).

    This is what supporting Linux looks like when you don’t build a bunch of custom scripts:
    http://www.opera.com/download/index.dml?platform=linux


    the concept that “too many distros, toolkits, api” turn off 3rd party developers is simply not supported.

    More distros + more windowing systems = more testing and support = more headaches for developers. That is a basic fact you can’t escape.


    developers dont target linux simply because of the market share

    Not true, even 1% is a significant market if is a single platform. As I pointed out earlier the iphone had better support from commercial game companies than Linux when it had .01% of the market.

    The iphone still doesn’t have 1% of the market and yet there is a waiting list to get your app approved.

  6. That’s true wisdom, brother Jason, what you’re saying at #3. I don’t think the best thing for Linux is wanting to be as succe$$ful as Windoze & Macz. The mighty mouse might be better off trying to do its own thing. But yes, that can be better streamlined, with more of cooperation and less of yet-another-newish-thingy. And also yes, brother Eric has a point: penguins should not automatically scare away from proprietary stuff. As long as the four freedoms are not threatened, just give it a go. If you can do the trick, they might even pay you for that…

  7. As far as “The first part means that 3rd party developers can expect to sell nowhere near as much of their software in Linux as they can on OS X or Windows.” goes…

    I think that’s a problem both for the Linux-users, the distros, and the 3rd-party developers. Let’s just give one example to show what I mean. An example that is not applicable to everyone but, from what I can tell, to an increasingly larger portion of Linux end-users (which is what I am).

    I use Linux exclusively. And I didn’t switch because of it’s free nature (in either use of the word free). I switched because it worked better for me than the competitive systems from Apple and Microsoft. That still holds true.

    However, if F/OSS can’t provide a program or solution to what I need to accomplish I am no stranger to using a free-to-use-but-closed-source (such as the flash-plugin) competitor or even a payed-for competitor (Maple is a good example, another is Turboprint).

    However, some non-F/OSS companies seems to take for granted that just because the product is there Linux-users will automatically buy it. Nero is the prime example of this, and the example I was getting to.

    Is Nero burning ROM a good program? Yes, as far as I can tell. It works fine but…it doesn’t provide (the Linux version that is) enough extra features compared to K3B or Brasero to be worth the money. All it can do can be had for free, so why should I shell out any money for it?

    What I’m trying to say is that Linux-users are used to good-quality free software in ways OSX or Windows users are not (even though they could be if they wanted to), so a 3rd party (commercial or otherwise) program needs to be something _really special_ for Linux-users to take notice of it, never mind paying for it.

  8. @eric
    iphone is backup by a major co operation with an extra large fanboy base capable and willing to by anything apple and that is why developers go to it and put up with all apple store policies

    with your logic, os X should have a bigger pie of computer games market because they have ones of everything but that is not the case ..their market share simply doesnt support the trouble of developing and porting games there

    chromium also has less than one percent market share and look at the impact it already has .. rekonq and future versions of firefox will all have its looks .. lack of market share can be helped by mind share a corporation behind a product but this isnt always the case.

    too many of everything on linux can be a factor, but it is not a major reason why 3rd party developers arent developing for linux ..

  9. Excellent point, @Jonas! And your example of Nero is an especially good one. So… this is a blessing and curse for Linux all around, right? Yay, we have lots of free “pretty good” programs! But because of that, we don’t get hardly any non-free “really great” programs?

  10. @mtz
    Yes, you listed some software. But no one said, that’s impossile to support (Desktop-)Linux. But it’s a lot more work to do, compared to OSX and Windows. And CSS has to do all the support by themselves. The farious fragmentations simply hamper support from CSS.

    And yes, all in all Linux ist hostile to CSS.

    Regarding chrome. It’s current marketshare is 3.58% – according to net applications. And it has a tremendous groth. Desktop-Linux on the other hand currently seems to stall.

    And the iPhone’s success is not based on fanboys. The iPhone just is a fabulous product, that for quite a while simply had no competition in many aspects.

  11. Regarding ALSA, there are packages of ALSA 1.0.21 with the MacBook Pro patches applied at my PPA: http://launchpad.net/~pgquiles/+archive/ppa (for Karmic)


  12. penguins should not automatically scare away from proprietary stuff. As long as the four freedoms are not threatened

    The four freedoms that Stallman declared as part of his war against proprietary software make the typical business model of selling software unprofitable.

    It’s only for a minority of software that you can give away the source and then make a profit through additional services or support.

    For most software when you give away the source you allow your competitors to undercut you out of the market. The vast majority of the work in developing software is in the creation of it, not in the distribution or support.

  13. @Jason,

    Yes, unfortunately I would say it’s both a blessing and a curse most of the time. The curse is mostly there, in light of Nero, that people might think Linux users are cheap-skates that just don’t want to pay for quality software without thinking _why_ they may not want to pay for the software in question. It’s applicable to other programs as well of course.

    Now, I don’t how many have paid for Nero Linux (to continue with that example) but I’m afraid that in the long run that the people in charge may jump to the conclusion that Linux-users as a whole are allergic to paying no matter what rather than seeing that they’re not offering enough of an incentive to upgrade from a free offer. I’m guessing here but I would think more people would be willing to pay for it if the Linux version was equal to the Windows (and Mac-version if one exists) version feature-wise.

    Or to continue: I have a legal license of Photoshop. Right now I’m using it using Wine. If a proper Linux-version was available, would I upgrade to it? I wouldn’t hesitate even for a second! Why? Because, if properly done, it would work better with the rest of my system than what’s currently the state.

    As far as the blessing goes…well, I guess it’s a matter of what you want. For productivity I’d take OpenOffice over MS Office any day of the week. Or Okular over Acrobat Reader. Or for entertainment: Amarok vs iTunes. iTunes is not even in the running of best all-around player (actually, iTunes is crap!).


  14. For productivity I’d take OpenOffice over MS Office any day of the week.

    But most businesses wouldn’t which presents a problem for Linux desktop adoption.

    iTunes is not even in the running of best all-around player (actually, iTunes is crap!).

    People don’t use iTunes as simply a player. In Windows you can use media player for that.

    They use to synch with their iphone/ipod, rent movies, buy music and games all with one account.

    Not having MSOffice and itunes hinders Linux desktop adoption. Linux needs to lose the anti-proprietary attitude if it wants to gain marketshare. Telling people to use free alternatives isn’t going to do it, especially when those alternatives don’t offer equal functionality.

  15. @Raymond,

    “But most businesses wouldn’t which presents a problem for Linux desktop adoption. ”

    I’m well aware of that. Still, it’s not as much of a requirement anymore (depending on the specific use-case naturally) as it once was. And I’ve seen more and more of businesses and government offices moving away from being dependent on MS-office. Whether it’s enough I leave up to others to decide. All I can say is that for me personally or the non-profit organization I’m part of, it hasn’t been a problem that I/we don’t have MS-office available.

    As far as itunes goes, yes. It has more to offer than a mere player – depending on where you are in the world. Where I am, it can buy music and that’s it. And I’m not sure that’s going to be much of an advantage for much longer.

    Still, I agree with your last paragraph. Linux does need to lose the anti-proprietary idea, which is what I was getting at. The way I see it, the OS should as far as possible be open and document-standards should be open. Specific programs (such as iTunes, Photoshop, Autocad and what not) does not necessarily have to be – as long as they’re easy to install, activate when necessary, and upgrade. Without conflicting with the core system.

    And yes, I think there’s a lot of work to be done there to make it easy. But I do think that the more traction Linux gather’s, the resistance of that will become less and less pronounced. What I’m trying to say is that I don’t think proprietary and F/OSS are necessarily enemies. They both serve their purposes and can complement one another.

  16. I think you should’ve mentioned that not only should we get more interest of software developers but also for hardware. It’s time hardware makers should start caring about Linux and shipping with “Linux compatible” stickers. With Google OS coming up and many Linux-based netbook distros it’s in the best interests for all of us.

    I think as a Linux user we’ve redefined exactly what “stable” is to me. To me, trunk is unstable, and snapshots are stable. To average computer users, “snapshots” is crazily unstable, and “last version released on the project’s website” is stable. For businesses, “last version released on the website” is unstable, and “3-4 releases before the latest release” is stable. I think software developers have to realise this mentality and really emphasise on bugreports as a responsibility to meet these perspectives.

    You should post a third blog post after debates in comments to finalise on a list of problems with Linux that everybody agrees with. Because right now everybody has different views of what the problems are.

  17. @Pau: Wow, cool! I’m a little wary of adding your PPA, though, since it looks like it’ll cause some problems with other things I have installed that you also keep in your PPA. =:( Any chance you could push your changes into the mactel PPA?

    @Jonas: Excellent points again. And I second that heartily. I have no interest in spending money for Nero because k3b does everything I need it to. However, I definitely did buy World of Goo and have bought several other things for Linux 1) to show my support of these vendors, and 2) because they’re genuinely useful.

    @Dion: Yes, absolutely. And truthfully, not having hardware vendor support is one of the biggest things keeping Desktop Linux from being 100% there. Good point. =:)

  18. @Raymond

    I wasn’t referring to these, but to Roosevelt’s (the originals)…

  19. “The first part means that 3rd party developers can expect to sell nowhere near as much of their software in Linux as they can on OS X or Windows”

    Jonas already pointed out that this is a flawed conclusion based on really good examples, however another approach would be to actually compare the situation on other platforms.

    No idea about OSX in this regard, but the whole end user experience on Windows is basically rooted in free as in gratis.
    Something is either Freeware, Shareware used with its limitations or cracked using key generators, or pirated copies.

    So not only do software companies have to compete against gratis software (could be Free Software), they also have to compete against software which is installed illegally, maybe even their own product.

    On Linux (if we take it as a placeholder for Free Software based operating systems), a company has to compete with the former (gratis software) but almost certainly not with the latter (mostly because their usual competitors are not targeting this marked yet).

    Putting the “everything else is gratis, users won’t pay” strawman into a discussion just shifts focus away from the main pain issue: deployment.

    Of course putting deployment into the superset category of development and mixing it with flawed examples like different audio output systems does not help either.
    (applications never access this part of the layer on any desktop platform, they work with multimedia APIs which abstract that away).

    Though that’s a mistake in Eric’s comment rather than your blog entry.

  20. It is possible to sell software to Linux users, they just have much higher standards- you need either an exceptional product (and you must be careful to ensure it stays that way, because your competition won’t sitting still) or you need to be selling a game. It’s quite possible (and can pay off handsomely for companies) to sell a Linux version of your game. Since many outside (and some inside) the community believe you can’t do it, any titles that do come out get a lot of press and attention from people.


  21. So not only do software companies have to compete against gratis software (could be Free Software), they also have to compete against software which is installed illegally, maybe even their own product.

    Most users don’t pirate, but more importantly piracy exists in Linux too. Selling proprietary software and dealing with piracy is still far more profitable than giving your source away and trying to sell services or support.


    Of course putting deployment into the superset category of development and mixing it with flawed examples like different audio output systems does not help either.

    It was all in the same category of Linux being a PITA for proprietary developers. You like many Linux advocates just want to downplay the harsh reality that developing and deploying proprietary software in Linux is far more work than in Windows or OSX.

    Developers not only have to wade through multiple and often conflicting sound apis but multiple distro and version differences that open source developers don’t have to deal with.

    Oh and in case you try to downplay the audio api mess:
    http://0pointer.de/blog/projects/guide-to-sound-apis.html

  22. Most users don’t pirate, but more importantly piracy exists in Linux too.

    Neither of us has any numbers, but I am quite sure that most users pirate, most commonly Windows and Microsoft Office, followed by “full versions” of software they got on their OEM installations.

    But I agree on piracy exisiting on any platform, the point was that a vendor competes against gratis everywhere, because it is the usual consumer mindset, not bound to a particular platform

    Selling proprietary software and dealing with piracy is still far more profitable than giving your source away and trying to sell services or support.

    Well, of course. No idea how you got into that context.
    This is about selling software. Wouldn’t make sense to discuss influence of gratis if your product is gratis as well, would it.

    It was all in the same category of Linux being a PITA for proprietary developers.

    Yes, but it shifts the focus and thus attentions for improvement away from more pressing issues like deployment.
    It dilutes the message into both directions, towards software vendors and platform providers, since the interpretation of the generic statement will be interpreted differently.

    You like many Linux advocates just want to downplay the harsh reality that developing and deploying proprietary software in Linux is far more work than in Windows or OSX.

    Not at all.
    While I am certainly an active member of a great Free Software community, I am also developing and deploying proprietary software on Linux, proprietary Unix and, at thankfully a decreasing rate, on Windows.

    Fortunately 21st century technologies allow that to be done by the same team using a single code base with only a few platform specialists doing the really low level platform specific bits.

    Experience shows that the major portion of development tasks do no longer have different difficulties on different platforms until you reach deployment.
    Which is why mixing these issues with highly academic ones just moves the focus away from any potential path to a solution.

    Developers not only have to wade through multiple and often conflicting sound apis but multiple distro and version differences that open source developers don’t have to deal with.

    Actually no. True, there are different sound APIs, but most of them are of no interest to application developers because they are way to low in the platform stack.

    Both Windows and OSX have APIs in their audio stacks where user space transitions to kernel space, but nobody expects applications developers to use them, so why should they do that on Linux?

    Application developers use APIs much higher in the platform stack, e.g. DirectShow on windows (probably CoreAudio on OS X). Because that’s what the codecs are available for.

    On Linux that would be GStreamer. It will work with the APIs in the lower parts of the stack, it will provide the codes, including proprietary ones.
    It has not had any incompatible changes in either API or ABI in years and its community (and the companies already building solutions on it) are committed to keep it that way.

    Discussions amongst developers of the audio stack as a whole, potential improvements, etc. certainly sound scary, which is why it is important to not mix it into the context of application development on Linux.
    Otherwise those discussions will have to be done in private in order to concentrate the outside view on topics relevant for that point of view.

  23. @Eric: Hah. Awesome link (http://0pointer.de/blog/projects/guide-to-sound-apis.html) =:)

    @Kevin: Excellent points. I guess I was confused. I thought pulseaudio is being pushed as the official sound API to write against everywhere?

  24. @eric: win32 binary compatibility?? oh, no…

    real story from today!!!: my company develops closed source win32 applications; today, we’ve just installed the latest windows xp updates, just to find out 3 hours later that a new vc runtime was pushed which breaks compatibility with all our existing clients which do not update also.

    Remember that WinSxs folder on Vista/Win7?

    Chris

  25. What really holds us back is our own disbelief and, to be frank, ignorance.

    Eric goes on about sound APIs. What I will grant is that sound performance is a clusterfuck on Linux distributions right at this moment. It’s getting sorted out, but various sound technology/API changes, particularly around PulseAudio, were handled exceptionally poorly by all involved. It was a typical “we are a distro, we know best, and now we’ll throw all caution to the wind and make this decision based on emotional/political factors” scene, something I really do hope the distros will mature out of.

    However, the sound API issue is a non-issue. How many KDE apps have need to be touched during this time period? Zero. Qt, assuming they were using Phonon? Zero. Now, whether the API works on the user’s computer is another matter, but that’s a job best left to something like Phonon, and is something that 3rd party devs can help ensure stays in good lick on all platforms. But that 3rd party code does not need to be affected by things lower in the stack.

    Ditto for HAL-cum-DeviceKit. Ditto for PolicyKit.

    3rd party devs should be looking above the bare OS APIs because that is not how to write applications that are portable or future proof. If your app is somehow performance critical, such as a high performance web server, you’ll end up writing to OS APIs. But anyone who says that’s what you need to do because you’re writing a phone app like Skype is giving you the wrong answers.

    Getting capture support into Phonon for things like webcams should be a priority for this reason.

    As for packaging for multiple distros, tools like Suse Studio start to make one wonder what all the fuss is about.

    When it comes to market share, we also get all down ourselves. We look at flikr or whatever our pet site is and the N. American markets. We buy the #s shoveled into the markets by companies with inherent bias. Remember Microsoft’s “93% of netbooks run windows” report? Turns out that our 7% is a lot more like 32% .. assuming we care about more than the market in the USA.

    Truth is that the mainstream small business and home user in North America and most of Europe is still wandering about in Windows land, with a good handful of them finding new love in Mac world. Turns out that those of us who create the bulk of Free software on the desktop travel in those circles. So what do we see? Yeah, it can be pretty disheartening.

    But if we bother to life our heads up and look around the word, turns out we’re doing pretty damn well. Better than Apple on most fronts if we care to measure globally, in fact.

    This is to be expected; the mainstream individual consumer in North America and Europe are more tech conservative (for a number of reasons) and highly influenced by marketing channels we don’t have reach into yet. The network effect this creates with the flikr’s and what not of the world impacts our inroads here as well.

    But here’s an odd thought: there are F/OSS, including KDE, apps that “do that flikr thing” and do it well. Who is approaching flikr and asking them to put that on their website? They won’t unless we ask, and even though they might say “no” if we do ask, it’s worth a shot.

    Stop misleading each other with narrow focus half-truths. We owe ourselves more than that given our efforts and our actual accomplishments to date. It may also give us the gumption to make serious inroads into the last major markets we haven’t made a dent in yet.

  26. @ Aaron.
    “Truth is that the mainstream small business and home user in North America and most of Europe is still wandering about in Windows land…”

    You’re right: indeed the world is wider than just North America. There is also Scotland, for example, and Belgium…. And also, reality is more than market shares. Whoever really counts these netbooks anyway? The main problem is still that John Doe, John Bull & Jan Rap (that’s the Dutchie) just never heard of Linux. Whenever I confess my faith in Linux, faces go blank.

    Yet, there is Hope! Quite a of these few blank faces, after a short pause, say ‘Aye, I think my nephew/niece said something about that’. Whenever I show them my actual KDE at work, they say, ‘That looks nice, and it isn’t difficult’. People are not against Linux, they just didn’t come across it, because it simply is not on the box they got from PC World or Curry’s, and Dell is hiding their few Ubuntu bits in the dungeons of their website.

    But now I get puzzled. Some say manufacturers don’t want to go Linux because FOSS is not generating profit and generally is a fuzz anyway. However, installing Linux instead of Windows on OEMs immediately pays off: it’s free. And about the fuzz: as long as drivers are made available, I — a common user — rarely had more fuzz with any Linux distro than I had with Win 95/98/2000/XP. The strong thing is: Linux forums may take some time to search, but usually you get an answer. Microsoft, or their ‘Knowledge Base’ too often didn’t answered when I got stuck.

    So I don’t think self flagellation or similar puritan excesses on behalf of the Linux community are helpful. Linuxees just need to overcome their tendency to FOSS dogmatism and effectively talk to these manufacturers, show them the wealth of support forums, the incredible ease of software repositories (as opposed to the endless and pointless searching across the whole world wide web if you want to keep your Windoze machine up to date) and challenge them to get these drivers organised. And of course it is all ‘at your own risk’. But is there any proprietary EULA that actually offers anything better?


  27. As for packaging for multiple distros, tools like Suse Studio start to make one wonder what all the fuss is about.

    Suse Studio is for building your own distro.

    But yes it is much easier to build your own distro than it is to build and support a proprietary application for a half dozen distros. It shouldn’t be that way.


    But if we bother to life our heads up and look around the word, turns out we’re doing pretty damn well.

    You’re in denial.

    Linux desktop usage peaked in 1998 and has been around 1% globally since 2000.
    http://gs.statcounter.com/#os-ww-monthly-200908-200910

    Here’s the real problem:
    1. People turn on a computer to run applications.

    2. Most applications are proprietary.

    3. Linux as an ecosystem is designed around open source and is a minefield for proprietary developers.

    Linux distros have software distribution systems that are built around open source. They are designed with the assumption that the developer will provide the source so each repository maintainer can make the needed adjustments to address differences between distros.

    Developing and deploying in Linux is a big mess for proprietary developers.

    This is what is looks like:
    http://www.opera.com/download/index.dml?platform=linux

    Until something like the app store is developed that works across Linux distros and welcomes proprietary software you can expect Linux to however around 1% on the desktop for another decade.

  28. @Eric: “Suse Studio is for building your own distro.”

    yes, you’re right. Studio is for full images. you can use the Suse build service with it, however, which just makes packages. for multiple OSes. visit http://en.opensuse.org/Build_Service

    “You’re in denial.”

    no, i have sales figures. knowledge is kind of the opposite of being in denial, don’t you think?

    “http://gs.statcounter.com/#os-ww-monthly-200908-200910”

    i guess you missed the part where i noted why Linux doesn’t show up in those kinds of stats.

    “Linux as an ecosystem is designed around open source and is a minefield for proprietary developers.”

    it isn’t as good as it could be, that’s completely true. but i also don’t think it’s a minefield, unless you pick your tools poorly. it’s not so much the API’s these days as it is the platform feature reliability, and that’s something distros do need to work on. that’s what’s bitten Skype which caused the original blog entry (of which this is the follow up to)

    even more challenging are the preconceived notions, some of which you’ve trotted out here yourself, about the status quo. e.g. that there are no good answers for how to build software for 98% of Linux desktop users.

    “This is what is looks like:
    http://www.opera.com/download/index.dml?platform=linux

    and others have given you examples where that isn’t what it looks like. you’re picking a very bad (or good, if you want to look at it in terms of platform coverage) example and trying to make it out to be the Way Things Are For Everyone. they aren’t.

    “Until something like the app store is developed that works across Linux distros and welcomes proprietary software you can expect Linux to however around 1% on the desktop for another decade.”

    i think that would certainly help, yes. and there’s probably a nice business model to be had there. why don’t you build it, Eric? that’d be a great project. you could clone Suse’s build service’s features, targeting the selection of distros that represent the overwhelming majority of desktop users in your target market(s), produce a slick client-side installer and offer access to it as well as distribution through a web service and your client app to proprietary vendors.

    there’s still the issue of user support, of course. but i’m sure you’d be up to the task of solving that one too.

    meanwhile, we’re busy, as a global community, doing things like helping ensure distros like Red Flag can continue to push multiple millions of units via retail a year (though not to your target markets, granted) or extending our reach into the global educational market such as in Brazil with the 50+ million students there.

    together, with your efforts to solve the bits that bug you the most (e.g. proprietary software availability) and our efforts to solve the bits that bug us the most, we can push through.

    bickering at each other doesn’t do much, though, does it? and that’s what i’d really like to see us get past.

    instead of arguing and blustering, trying to find more factoids that can be made to back one’s pet position, let’s actually fix things.

    are you up for that?

  29. Eric, you repeat these two points over and over like some kind of mantra:

    1. Linux as an ecosystem is designed around open source and is a minefield for proprietary developers.

    2. Developing and deploying in Linux is a big mess for proprietary developers.

    In my experience, neither of these is true. I work for a third party developer and we ship applications on Linux, Mac OS, Solaris, and Windows. The truth is that we have to spend as much time dealing with quirks and constant updates of the proprietary OSes as we do on Linux. Do we care one jot that Linux is open source? No.

    Hmm.. actually that’s not quite true. In addition to developing our desktop products on Linux (we use Qt), we build all of our hardware products with Linux (no extra cost or licensing hassles, thank you very much), and we can’t forget that our internal servers run it as well. So I guess we really do care that Linux is FLOSS!

    Yes, there is much that can be improved to make life easier for Linux users and developers, but the situation is hardly as dire as Eric claims.

  30. @Aaron
    Suse build service is nice. But only for OSS. It does not help CSS.

  31. @Kevin: Excellent points. I guess I was confused. I thought pulseaudio is being pushed as the official sound API to write against everywhere?

    That’s one of the problems of putting “internal” discussions into a different context, e.g. application development.

    It is something that is below the layer an application developer would usually work with unless the application uses its own multimedia framework, e.g. Apple’s QuickTime player.

    Communication among developers working on infrastructure is usually available for proprietary platforms, so blogs like the one from Lennard are easily misunderstandable as communication with developers using the infrastructure.

    Bringing them up in the context of issues third party developers are facing makes it harder to see these issues, they kind of get shadowed by the more serious sounding internal ones.

    And it makes it next to impossible to motivate people to work on solutions to these problems because they know any work will be ignored and academic problems reiterated instead.

  32. Linux is a minefield, but not only for propietary software developers, but for free software ones too. Ever wondered why Firefox and OpenOffice work better on Windows?

    Linux is a complete, irreparable mess the moment you go further than the command line. That’s why server and embedded Linux is doing just fine, and desktop Linux is nowhere to be found.

    I don’t expect it to ever change, only more and more denial for many years to come.

  33. @Roy,

    Work better in what way? Not that I’ve noticed anyway, so please elaborate.

  34. @Birdy: “Suse build service is nice. But only for OSS. It does not help CSS.”

    yes, i know. note that i suggested Eric takes a proprietary-friendly clone of it on as a project. but the Suse build services shows it’s doable in a practical, user friendly way. which makes Eric’s bellyaching just that: bellyaching.

    step up or get the hell out of the way, right?

    @Roy Orson: just because you say something as if it were factual doesn’t make it a fact. you really need to back up your assertions. if you have a good point, you should be able to explain it and then perhaps something will get done about it. otherwise, it’s useless noise that serves only to discourage and reinforce a self-defeating attitude.


  35. In my experience, neither of these is true. I work for a third party developer and we ship applications on Linux, Mac OS, Solaris, and Windows. The truth is that we have to spend as much time dealing with quirks and constant updates of the proprietary OSes as we do on Linux. Do we care one jot that Linux is open source? No.

    Is it a GUI application that has to be deployed on multiple distros? If so then you are lying.

    Do you really think it is just as easy for Opera to maintain a dozen packages for a dozen different distros (plus extra packages for multiple versions of those distros) than it is for them to maintain a single executable for 7/Vista/XP? Why can’t Opera provide even a single binary for Ubuntu 8.04 and 9.10?

    You guys can choose to live in denial but don’t expect everyone else to.

    Ian Murdock of debIAN
    Unless an application is included with your Linux distribution of choice, installing that application on Linux is a nightmare compared to Windows.

    http://ianmurdock.com/linux/software-installation-on-linux-today-it-sucks-part-1/

    Firefox lead Ben Goodger on porting to Linux
    http://linux.slashdot.org/story/09/05/30/1740205/Harsh-Words-From-Google-On-Linux-Development?art

    The attitude of the kernel devs is fuck proprietary drivers, even if that is what most hardware companies want to release.

    The attitude of the distro maintainers is fuck proprietary software, even if that is what most developers want to release.

    For years Linux advocates have said to just give it time. Just give it time, hardware companies will relent and developers will either open their source code or a gpl alternative will develop eventually.

    Well we have given it time. Linux was more popular in 1998. It’s been flatlined at 1% since 2000.

    The fuck proprietary software attitude is not working. More people would rather buy a $1700 macbook than run Linux.

    But I’ll let you get back to your little denial tea party where you can blame OEMS or M$ for the failure of Linux on the desktop.


  36. “http://gs.statcounter.com/#os-ww-monthly-200908-200910″

    i guess you missed the part where i noted why Linux doesn’t show up in those kinds of stats.

    No I think you missed that Linux is the red line at the bottom with .068%.


    why don’t you build it, Eric? that’d be a great project. you could clone Suse’s build service’s features, targeting the selection of distros that represent the overwhelming majority of desktop users in your target market(s).

    Thanks for proving how unappealing Linux is to proprietary developers by suggesting that I build my own distribution service. That’s almost as funny as suggesting that I build my own distro (which I have been told before).

    Anyways the first distro who builds an app store will get the FOSS army up their ass over violating Stallman’s newspeak definition of freedom. So even if a third-party built such a system there is no guarantee that the major distros would support it. Building it without their support would result in lousy integration.

    So expect the current clusterfuck situation to continue. I’ll stick with Windows and OSX.

  37. “So even if a third-party built such a system there is no guarantee that the major distros would support it”

    ah, i should also note that it probably wouldn’t matter if major distros supported it. if there is a market for the software on Linux, people will happily buy it from your store without caring if their distro supports it.

    just like people are happy to buy software from Adobe even if Microsoft doesn’t help in the transaction or the upkeep of that software.


  38. ah, i should also note that it probably wouldn’t matter if major distros supported it. if there is a market for the software on Linux, people will happily buy it from your store without caring if their distro supports it.

    just like people are happy to buy software from Adobe even if Microsoft doesn’t help in the transaction or the upkeep of that software.

    When you buy software from Adobe it still works with the Windows program manager. Microsoft still helps in making sure the program integrates with the system.

    Without distro support you wouldn’t have that same level of integration.

  39. Pingback: Why Isn’t Desktop Linux “There” Yet? - Kasperian Moving Parts

  40. Try FreeBSD, or if you want software completely free from closed source code try OpenBSD. BSD has better wifi driver from Linux because if BSD developer can’t do it right, they don’t release it. If you can’t find similar app port in BSD, BSD can run Linux app.

    I haven’t left, nor do I plan to, Desktop Linux (I know, who cares), but I think it’s important to think about what might be broken in our development model and figure out how to fix it,
    I think the problem is:
    http://www.roughlydrafted.com/0506.linuxmyth4.1.html
    “Worse is Better – Linux is designed to be good enough to get the job done when the job could be most anything. Linus Torvalds doesn’t claim that Linux is a perfect architecture and avoids comparisons about having the best design. Linux’ code quality is not unquestioned.”

    http://apcmag.com/interview_with_con_kolivas_part_1_computing_is_boring.htm
    “Of course it did. There were so many subsystems being repeatedly rewritten that there was never-ending breakage. And rewriting working subsystems and breaking them is far more important than something that might improve the desktop right?”

    “Even worse than that, while I obviously like to see Linux run on 1024 CPUs and 1000 hard drives, I loathe the fact that to implement that we have to kill performance on the desktop. What’s that? Kill performance? Yes, that’s what I mean”

    http://www.forbes.com/2005/06/16/linux-bsd-unix-cz_dl_0616theo.html
    “Linux has never been about quality. There are so many parts of the system that are just these cheap little hacks, and it happens to run.”

    Here Mac OSX vs Linux:
    http://www.roughlydrafted.com/RD/RDM.Tech.Q1.07/8510F3F6-AA84-4324-8882-1319FB939E8A.html

Leave a Reply

Required fields are marked *.