What does the free desktop need to grow in market share?

Saw a story on OpenOffice 3 today which reminded me of a question I been asking myself recently. What does the free desktop need to grow in market share?

Up to this point I guess I my thinking about the free desktop (grouping GNOME, KDE, XFCE etc. as one) and its growth has mostly been about seeing it as a dam filling up. The mass migrations from Windows would be trickling in slowly until at some point we have added enough features and polish to the free desktop for the dam to break. Kinda like how linux in the server space lived many years without a lot of adoption outside academia and or specific fields before suddenly becoming an ‘overnight’ success.

But at this point I am not so sure anymore. I mean is what holding us back from rapid gains in marketshare really just better MS Word import in OpenOffice? Or better support for exchange servers in Evolution? Or better drawing tools in Inkscape and Gimp? Or better support for muxing Quicktime files in GStreamer? Or improved ways of embedding a blingy clock widget into the desktop background? Or just adding an application that can do XYZ? Or is it the lack of a good driver for hardware ZYX? Sure these questions are part of the answer, but I can’ t help but wonder if they are a smaller part than I have given them credit for so far.

I have sometimes seen the lack of games being mentioned, but the Windows game market is suffering terribly these days, caught between piracy and console dominance. So if people care about games I think they probably got themselves a Wii, PS3 or Xbox360 to satisfy that need. So I can’t see lack of game support as being the tipping point either.

Not that there isn’t progress made. There are good migration stories out there from the major Linux vendors and PC makers like Dell do seem to try to offer better sales support for Linux desktop systems. But I can’t help but feel that we might be missing something in terms of understanding what needs to happen for the market share to grow more rapidly. And if we don’t diagnose the issue we will not be able to resolve it.

48 thoughts on “What does the free desktop need to grow in market share?”

  1. Its all about comfort. While this may not be a big problem for new users or people already brought up on Linux, to gain marketshare for existing Windows users, the linux desktop experience has to be massively better than the equivalent available on Windows, otherwise the inertia of being comfortable is too large to overcome.

    If it is the same or similar, why bother to move to something where there is no improvement, but many things are oh so slightly different.

    would class myself as a person who fits in with the above problem – for me, Linux is generally a better desktop. It does everything I need, the majority of it slightly better than what Windows does.

    However, I have not made the jump as I am comfortable with Windows, and there is nothing marginally better about Linux. Except maybe wobbly windows.

  2. What “You” said. It’s a matter of comfort. And those damn Word documents. And lack of a killer app.

  3. Perhaps this might seem obvious, but I think it largely depends upon which audience we intend on attracting. For some workloads (basic office productivity), it seems to me like the major distributions are already quite adequate although not too far beyond adequate. To attract this sector, to which money is often no object and absolutely compatibility a must, we must offer something truly unique. There must be a positive draw for business users. There is no denying that cost of retooling an existing Windows setup is quite high. Regardless, I would say that on this front, we have made very good progress.

    When it comes to attracting casual users such as my fellow college students, we need to take a close look at usage patterns. Most students, for instance, do very little other than basic word processing, instant messaging, and web browsing, specifically YouTube. In my evaluation, we’re doing well on each of those counts besides than the last.

    While it’s quite frustrating to think that we are being held back by something as simple as adequate Flash support, but it seems to largely be the reality. When I (on my 64-bit machine) try to show a YouTube video to some friends and my browser freezes, Linux looks bad. When I have to then open a terminal, and enter some arcane-looking commands to kill the browser, Linux looks very bad. When I launch my browser again and have it happen again, Linux looks downright primitive.

    Unfortunately, it seems like many of these trivialities and annoyances such as lacking Flash-support score high on the user-relevance scale. For casual users (who I would otherwise say is a natural audience for Linux), this means Linux is not an option.

    It seems to me that there are two ways that this problem needs to be attacked. First, we need to work on improving features that common users actually use (e.g. swfdec). Secondly, and perhaps a much more difficult challenge, we need to promote those features that differentiate Linux from its competition. Linux cannot be viewed as just an alternative to Windows. Apple is a perfect model here. I think that we need to be drawing attention to _high quality_, well-integrated open-source software that allows you to use your computer in new ways. The high-quality qualifier is a must; pushing buggy, bit-rotting or unpolished software, no matter how useful, unique, or innovative, cheapens the desktop. Jokosher is a excellent example of the sort of software that could help distinguish the Linux desktop.

    Anyways, that’s just my 2*10^-8 cents.

  4. We need more applications with better integration, performance, features, and usability.

    Having 75% of the features, 25% of the performance and 50% of the usability of an existing application isn’t good enough to make people take the leap. (Looking at open office).

  5. The little features _do_ matter. A polishing trend with attention to little details improves the quality of any project as a whole.

    It’s like going from “less than the sum of its parts” to “more than the sum of its parts.” The little touches help fuse the cohesive experience into a synergistically effective whole.

    It’s like going from a music player that says “Please install these tools so I can do the job” — or even worse — “The required tools to do the job you request are not available” to “Let me setup the necessary requirements for you, and you may change the settings later if you prefer.” (in terms of the music player analogy, I’m thinking of things like album art support, replay gain functionality, and CD-burning — all the little things that have to be pieced together _well_ to make a hypothetically good multimedia library app).

  6. People are running out of reasons not to use a Free desktop, but they’re not really gaining reasons to use a Free desktop. I don’t really know what those reasons might be, unfortunately :(

  7. Gnome is a lot more straightforward and usable than windows xp and vista are out of comparizon with the most confusing interface ever.
    For media the Amarok is far more better tool that media player and very usable also.
    The problems I face with linux are most of the times hardware issues, it takes time to support new hardware and some examples are, 801.11n wireless cards, my pinacle tv adapter, poor support for bluetooth devices (i threw away my logitech dinovo cause the bluetooth stack was problematic in linux).
    Also the gaming is a big issue cause pc gamers cannot be easily satisfied with a ps3 for example.
    I like linux for many reasons and I ignore these problems but I would be happier without these problems, but most users cannot tolerate these problems that’s why the market share is small.

  8. The fact is that the best open source tools on Linux are multiplatform. So people can use them on windows and on osx. While the rest of the software available looks like an unprofessional/uncompleted/alfa-thing.
    Why should they switch to a completely FREE Desktop? Most of the people don’t care if it is “Free as Freedom” or not, they want something that helps to get their life easier. Most of the Linux users don’t care either about the “freedom” or they wouldn’t install nvidia drivers, a blob of codecs and the adobe flash player. What about the developers? As far as I know reading the planets a lot of them use a Mac as main desktop, if they don’t use it why the grandma should do it?

    On Windows you can use the best desktop tools available on Linux plus you can have the best proprietary programs for FREE, Free as Pirated Beer.

    I use only Linux but because it’s my choice. But I understand why a lot of people aren’t moving to Linux. It’s harder and there is less choice.

    Now comes my stupid tip, I’m sure a lot of people won’t like:

    I think the only way to attract users is to have some cool programs that work only on Linux and nowhere else. I don’t know if it is possible to write an open source program this way, I’m not a programmer. But if it is not… it’s better to write some very cool closed source programs that work only for Linux and release them under GPL only when finally Linux has gained several points in the market.

  9. I think you all are missing the point: yours idea are all good but cover only the 10% of the task (gain market share) but the remaining 90% is always untouched. There is only one company that “owns” the global market of computer and this company will always make all it can (and something more) to let things remain the same. In my country (Italy), computer with linux preinstalled came to reality only this summer whith netboobks and two models of very underpowered laptops: so what can you expect? The race to the desktop has ONLY just begun….

  10. The mistake that I think is being made is an over aggressive desire of GNOME developers to attract home users. It’s a segment not in need of a new OS. Home users want to surf, play their videos (and have them all work), play DVDs and use iTunes. All of which is not strong on Linux.

    Growth can only come in business and government, but that market is not being strongly considered; and seems not interesting to the developers. In the case of government, there is a HUGE potential for market share growth. Yet another media player isn’t going to help; we need stronger Evolution enterprise features; stronger OpenOffice features; stronger document management and better ability to connect to backend systems (gnome-terminal, 3270, Rdesktop, Citrix, NX). It’s 2008 and Evolution still doesn’t retract a meeting–those things are just killing the ability for growth.

    Regards

  11. Although I’m exclusively on Linux, for Joe user, I’ll have to agree with “You”.

    It’s more about habit than anything, and the majority of users are not interested in computers, and will user whatever thrown upon them at work, university etc.

    For this to turn, Linux on the corporate desktop needs attention. That way, the universities will follow, and hopefully (everyone) else.

    Going the other way ( like, incidentally, I am) will always yield extra work, and is of no interest to people who are not free software advocates, or ‘tinkerers’.

    At least, that’s my opinion.

  12. Personally I feel that “just works” stability has not been as good in 2008 as I had expected it to be. Previously, each passing year would build on what “just works” on the Linux Desktop (whatever the distro).
    Partly due to the big change in the Linux audio space (most pronounced with the Flash problem) and some other big changes, 2008 distros have not been as hassle-free to new (or experienced) users.
    Hopefully 2009 will not see such disruptive changes.

  13. i am more concerned about the lack of professional software.

    * there is no CAD program for architects
    * there is no video editor for film makers
    * there is no photo solution for photographers

    what counts in the end are applications that are data/cpu intensive,
    other tasks are more and more provided as webapps.

  14. If we want the world to embrace free software, we have to make it beautiful. I’m not talking about inner beauty, not elegance, not ideological purity… pure, unadulterated, raw, visceral, lustful, shallow, skin deep beauty.

    We have to make it gorgeous. We have to make it easy on the eye. We have to make it take your friend’s breath away.

  15. I disagree with almost every post so far. I am an administrator of small and medium sized businesses. We choose Windows nearly exclusively, despite me running Linux on my desktop. We choose Windows for most servers too. When I say we, I actually me “ME”. I make these decisions.

    I find the problem to be completely misunderstood by every free softawre enthusiast I come into contact with. Most of these people have never been in charge of an environment such as mine. An actual business. With actual real live human beings. Where the focus of the business is not on software itself. Where the business is not full of people desiring to work on software, but full of people, trying to get business tasks done. I have real business tasks to get done. I need an email server which is nearly 100% hands off, as Exchange is. I need a domain authentication infrastructure that is nearly hands off, as Active Directory is. I need desktops that are nearly hands off, and that work seamlessly with all of the above. I need DHCP, hands off. DNS, hands off. I do not need to learn how to configure bind. I do not need to learn how to configure http.conf. I do not need to learn how to configure slapd.conf. I need a server operating system where I can put the CD in, walk through 4 simple questions about my environment, and it sets *all of these* up for me. In a consistant, supported manner, designed for my business needs. I need the flexibility to expand the business without worrying about the technology. I need an email server which automatically authenticates against my infrastructure. I need the database server to do the same. The web server, the same. I run a successful business with all of this infrastructure with one real IT staff member, me. I need a system that quite simply is polished, so I don’t have to do work. The work is already done.

    No Linux distro comes close to this. None. Novells’ Open Workgroup stuff comes the closest, but it is NOWHERE near Microsoft’s offering.

    There are hundreds of thousands like me. I do not ‘care’ about the software being free. The server software I was just speaking of, which does DNS, DHCP, Kerberos, LDAP, automatic deployment of desktops and has UIs for every bit of this, costs $1000. That’s it. I make that much in salary in less than a week. I might not be able to edit the code, but I have no intentions to in the first place. It works fine without me editing it.

    Free software will not win my business until you can compete in that space, as MS did, and continues to do. If that means free softare will never win, then it won’t. That said, a radical shift of focus away from desktop gadgetry and to actual valuable process automation and polish would do it for me. But nobody can convince the masses of free software developers to do this. Nobody WANTS to do it. As long as the ecosystem of free software development consists of people working on what they WANT to, it WILL NOT HAPPEN. Period.

    Seriously, we have LDAP servers. We have DNS servers. We have mail servers, some pretty cool ones. We have all the peices. But we have nobody interested in packaging those pieces into a coherent set that can be deployed by people in my position. For every person who likes editing configuration files by hands, 99 others do not. Free software is addressing that 1%, and it shows in the market share.

    Microsoft worked on this from the opposite direction. Their first priority was getting businesses using their software. And they succeeded. They did everything required to solve real business problems with their software. To address the HUGE segment of the market who wants these solutions, and has no intention of learning how to hand edit configuration files.

    And they succeeded. They got the market. We can harp on them as much as we might want for making ‘technically inferior products’, but they won. A long time ago. What’s the goal if not winning?

    So, I highly value free/open source software. I love having the code available to edit. That *is* a bonus. But it isn’t a pro that outweighs the significant pros on the other side of the fence. TCO is real. It matters.

  16. 3rd party software distribution. Creating something valuable that people will pay for is more or less impossible. Like the troubles of glom and abiword who are at the mercy of distributers ‘sorry business is closed you can’t add your software’. And that’s just free stuff!
    Rant follows…..
    I can’t finish my work tomorrow and get it out there for users. I have to wait for next distribution update and when users update to that. Same story with hardware, just less predictable with the crappy kernel release policies.
    “Oooh we ship more drivers than windows!!!!” That doesn’t matter, it matters if the hardware i buy today works with my computer. It’s impossible to create a driver that will work for the next 5 years because linux internals always change because engineers have convinced themselfes that stable API/ABI is the root of all evil (I claim it’s because it’s boring to maintain and they are an arrogant bunch who finally found something they can work on without anyone telling them that they should do this boring work). Ofcourse new features will be finished slower if the internals of the kernel has to be stable for a certain amount, but if there are no users for the fancy features…. then does it matter?

    Ofcourse it’s not the whole story, first to market also helps and more….

  17. Can’t say anything about the “enterprise sysadmin” view (except maybe that upgrading the RHEL systems at work to a new release is a huge task, hence admins only do that every few years, hence our workstations are consistently outdated compared to current “Linux desktops”, hence questions like “why doesn’t this stupid thing play my audio cd/MP3/Video/DVD/whatever” are still common even if you as Gstreamer dev might think this is long solved).

    Aaaanyay :-) I fully agree with others that to reach home users, the free desktop needs something _exceptional_ so people just “want it” and even accept that they have to relearn from Windows to Gnome or KDE. But so far, I don’t see a reason for people to switch to free desktops. Just being as good as Windows or MacOS is not good enough (not that Gnome or KDE are actually as good as MacOS, unfortunately).

    With regard to Marcos “write some very cool closed source programs that work only for Linux”: I don’t think it has to be closed source… This exceptional feature can also be something that just can’t be ported easily to other systems. Stuff like Compiz for example (only that it is several years too late to make people go “wow” :) , or maybe PulseAudios spatial sounds (only that it’s probably too difficult to effectively demonstrate its awesomeness), or maybe Federicos “document-centric” ideas? Maybe this exceptional feature can even be “the desktop where everything just works”, but seeing how many minor and subtle bugs there are still in current free desktops, I suppose this feature would require way more manpower than is available so far, and is way more difficult than some genius idea that is turned into a dam-breaking feature.

  18. As Havoc said many times before, first choose your audience.

    For example, for the home desktop, it’s lack of a killer app, as said before. Linux doesn’t cost much, but also doesn’t add much. OSX is a killer and apparently many people are willing to go through migration pains to enjoy it (and it’s worth it!), Windows is already there. Why change? Linux just isn’t better. The gains are not worth the migration pains. The out-of-the-box experience (which is not the one that you get after adding millions of extra repositories, using SVN/trunk code and installing all kind of extra plugins) needs to be sufficiently smooth to get people to stick and there needs to be a killer feature (e.g. a killer app) to get people to change. One you’re there, you’re where OSX is now, roughly.

    For example, for the business desktop in large-deployment environments (>1000users), it’s lack of comfort of using Linux, as Jerome just explained.

    For small businesses (2 programmers and a secretary), “Linux” may in fact work already.

    (Etc.)

  19. My point is, in short : for Linux to get widespread adoption, target the enterprise. When they adopt Linux, most people will want it home as well, to avoid learning two OSes.

    What I would like, though, is a slim user (as opposed to enterprise) desktop with impressive features, although that won’t get Linux any closer to “world domination”.

  20. Linux is still pretty limited when it comes to quality content creation applications. As soon as i go to use Linux for more than email, music, video and web browser i have this pit in my stomach that things are going to start crashing or that i will have to deal with a perplexing user interface that makes doing simple things hard. Don’t get me wrong consuming content on Ubuntu is much better than on windows. Gstreamer/Multimedia is amazing! I have more codec troubles on windows (videos without sound, songs without metadata) than on Ubuntu! Its just when it comes to content creation, the offerings are fairly limited and/or buggy and/or ugly and cumbersome.

    I also love Gnome/Ubuntu because i like the fact that it is built by a community, and i like watching the technologies take shape, but at the same time i recognise that unless companies get involved and drive application development, we simply are not going to see the feature complete, refined and crash free systems that you expect from the best of other platforms.

    So in summary. Linux is good if you are a content consumer, but its more limited if you are a content creator.

  21. I think bugs are the main reason people don’t make the switch to Linux. It is very easy to break your system. Although Windows and OsX have their share of bugs, they are not as many as in Linux. And people need something that’s easy and works.

    People should test their code more before pushing to stable branches. And people should also spend more time on bugzilla, trying to fix bugs.

  22. I used Windows since 1991 and I switch to GNU/Linux (Ubuntu) a year ago. Tools we use for our work include databases, STAT software, Web deploying etc. We faced all the possibles problems with proprietaries software and windows itself. (To the point to our computer was locked, so we could not recover our information with proprietary expensive software). It is the adoption of Linux easy? Not yet. Why? because the learning takes some time. But at the end you have the best OS for any task at this moment. Learning time is vital for grandma, so she don’t get bored and quit. Ubuntu slogan is the key “Linux for human beings” . What we can do? work, work and work for Linux whatever front you want. I’m not an IT but I develop , and any software that we will make it will licensed GPL3.

  23. I think everything that is holding back Linux from larger market share has been mentioned in the above ~24 comments. There isn’t some piece of the puzzle we just aren’t seeing. The problem is that the above ideas on how to improve Linux are hard to accomplish and take time, money and effort.

    Mozilla/FireFox is a much better example to hold up to the light for examination than how Linux took over the server market because it’s a lot less complex. How long has Mozilla better than IE? I would say somewhere in the 0.9x time frame it was markedly better but it got no traction. That might be debatable so lets just skip right to FF3. You would be hard pressed to find anyone that didn’t agree that FF is multiple times better than IE and yet it holds a third of the market share and will probably hold there for some time. The Linux desktop has to be better than the Windows/Mac desktop in the same way FF is better than IE times 10. This is because the desktop is a much more emotional change than just switching browsers.

  24. I wanted to make this point separate since it’s just a personal pet peeve of mine and doesn’t really advance the core of your post.

    Windows gaming is NOT in dire straights. WOW is printing money. Steam is going strong. Warhammer hit 750,000 customers in less than a month. All the games I mention do NOT show up in the revenue breakdowns you see by platform. This is because they are subscription services and aren’t really comparable to most console games sales. WOW alone has so many subscribers it’s the equlivant to releasing a new version of Halo every month from a revenue perspective. Warhammer alone is pulling in over 11M/month just with their tiny 750k customers. That pales in comparison to Halo3’s one month figure of 300M, but that 300M is a one time shot and they have to spend another 2-3 years to make another while Warhammer is steadily increasing their revenue each month.

    Then look at games like Spore, Diablo3 and Crysis which make up 14% of the overall game market. No one knows exactly much money is made in subscription games so a true % of market can’t be derived.

  25. I love Linux. I love free Unix systems in general. I’ve been using free and open software for almost ten years and I’m still constantly amazed and inspired by it, but I personally don’t quite, fully understand the obsession some people have with pushing software such as Linux as the be-all-end-all solution for every computer user and scenario in existence. We need diversity, competition and the ‘bad guys’ to make things interesting and challenging. The right tool for the job etc. etc.

    I often wonder how many ‘average’ ‘do-what-I-want-without-frigging-around’ home Windows users *would* start using something like GNOME+Linux even if there *was* a collection of ‘killer apps’. No matter what benefits it offers, it’s still ‘something different’ that doesn’t work the way they expect.

    For me, it’s much more satisfying to be able to explain the basic ideas behind free/open source development to a non-geek friend and have them understand and appreciate what it’s all about. Especially when it’s someone who might have used and enjoyed free software at some point (maybe even without knowing).

  26. Of course, there’s a huge part of “confort”. Don’t move, don’t change, ever.

    But, in the industry, there’s still one huge technical factor : MS Office.
    – Excell sheets with macro that doesn’t work in OpenOffice
    – No Visio equivalent (that should support the Visio format)
    – OpenOffice is… ugly

    The last one seems a personnal opinion, it is not. I’m astonished by the number of people who could use Linux, who can use Linux, who have a lot of advantages in using Linux but who keep telling me : “Yeah, I know, Windows sucks. But, honnestly, I cannot work with OpenOffice, the icon are so ugly.”

  27. You have the part about Windows, games and piracy the wrong way: why someone would bother with a console then he can use a Windows PC and have easy access to tons of “pirated” games? From this point of view, Windows is *a lot* more convenient than other alternative.

  28. It’s actually very simple: the vast majority of windows users are locked into Word (not Excel, PowerPoint and definitely not Photoshop, CAD or any other software). And they are not locked into Word because there is no good import into OpenOffice, but because there is no perfect import into OpenOffice.
    That’s it. That is all.
    And Microsoft and Apple are very well aware of this.
    That’s the reason Microsoft has put such a large effort into corrupting ISO…


  29. I often wonder how many ‘average’ ‘do-what-I-want-without-frigging-around’ home Windows users *would* start using something like GNOME+Linux even if there *was* a collection of ‘killer apps’. No matter what benefits it offers, it’s still ’something different’ that doesn’t work the way they expect.

    The answer is very simple : They will not ever switch on their own, simply because they then will have to learn two OSes – Windows at work and Linux at home.

    That’s why the enterprise should be targeted.

  30. In the following I would like to argue for something that I normally do not argue for. In my experience with Linux, and particularly GNOME, what I think is *missing* is a raft of market-specific applications and spins of GNOME-desktop targeted at those specific markets.

    Over the years there have been many, primarily debian-based, Linux distributions which are highly catered for specific market segments-ie. audio production, mythtv, etc. Additionaly we have Linux distributions targeted at the so-called “enterprise”. And of course we have things like edubuntu.

    What I would like to see is GNOME being spun for specific market segments-ie. the problem with distributions targeted at specific market is that these distributions include *everything* available in the Linux world for that specific application, yet what we need is are distributions where GNOME is tailored for the specific market and included GNOME specific applications to address missing functionality-and only when we attempt to do this are we going to really see what is actually missing.

    Here is a small list of what I have in mind:
    GNOME-Academia
    GNOME-Office
    GNOME-Admin
    GNOME-thinclient
    GNOME-Publisher
    GNOME-Business-manager

    To further illustrate this I will focus first on GNOME-Academia. Imagine a GNOME focused distribution which targeted academia. Right now Linux fails in academia use because of simple things like the lack of SPSS. R is free software-with some work to tie in nice cairo-based graphing applications and a GNOME front end to R such a GNOME-Academia version could trully *replace* hundreds of thousands of existing Windows installations which only exist to offer SPSS to students. If in addition to this GNOME specific software was availalbe to facilitate citations, to track quotes and faciliate research, GNOME-Academia could become the attractive software for University students who do most of their university related work using computers. The primary question would be -what do university researchers need in the way of functionality ? what would it take to provide a custom-tailored desktop which not only matches their needs but exceeds it by offering simple intuitive integrated functionality. We already have some software offering some of this functionality-but it is not integrated, does not have a GNOME face(ie. is not simple and usable and clean which are hallmarks of GNOME programs.) If there were an application which was tailored for document production which enabled one to tie a smallish set of custom-tailored applications together for the production of academic work we could make their lives significantly easier. So an application which would be able to pull data from a word process(OO,Abiword), from spreadsheet (OO, GNUMERIC), from stats data R, from custom databases (GLOM), from mathematical/scientific formulas (Latex entry), graphing programs, and bibliographic/citation systems. I envision one app as being a kind of shell around 5-8 smaller applications where everything can be integrated in document publication. Much of this work has already been done, much code already exists which could be ported over to GNOME, but if one would take this as a target and focus on creating a holistic GNOME environment for this specific market segment, we could have an offering equivalent to what Macintosh once held in the profession publishing world. If we also had tools which would allow specific academic fields and even specific faculities to publish there templates for journal articles, citation templates, artwork etc. (be they latex, or OpenOffice, or Abiword) to community/project based portals then indidivuals and institutions using this could tie their own creativity back into the project-creating strong identification and strong feedback systems.

    In the same vein a version of GNOME-Business Manager could be tailored which had a group of GNOME applications which specifically redressed the needs of managers and small business owners. Things like accounts recievable, payroll, inventory, contact management, project management etc. The idea being yet again- a GNOME desktop absolutely custom-tailored to the exact needs of this market. Concievably it might be wise to tackle certain applications first as web applications-drawing from a larger crowd of potential contributors due to wider cross-platform availability with the goal of then taking the result of such widespread colaboration and churning out smaller apps which properly integrated into a work flow offer this GNOME desktop a unique advantage. Here I am thinking of things like tax programs- it is hard to find coders who are tax experts-use a web-based tax application project to gather the data from a really large cross-platform population and distill the thus collected data into specific GNOME desktop applications, and tie these into an integrated workflow. The goal would be that a small business owner would look at this desktop and say-YES! everything I need to do my work is here and just look at the work flow!-I use this app to do this and that app to do that and at each step this contextual knowledge is available for these other apps which makes my work so easy. Each society and each small business have specific regulatory regimes (taxes, laws, codes etc.) Custom tailored software which ties into these specific contexts could really empower people and if it was possible to allow these users to codify such context specifics and render such accessible through these applications and then share these with others of the same profession or residing in the same society identities and communities could be built around this.

    I focus on GNOME specific apps, because the key to this stuff is contextual integration in established workflows. In some cases software may already exist which does certain parts of this really well but is not GNOME specific-then the question becomes can we create a small set of function-specific applications which are custom tailored to integrate with this non-GNOME software(perhaps Chandler would make the best PIM for a manager-how could GNOME apps be tailored to tie into Chandler contextually…).

    Of course there would be significant overlap in the functionality and applications available on these different spins-but one could market these spins for specific market segments. And of course such projects would require writing much in the way of new software-but most of the infrastructure(libs) already exists-what is needed is small graphical applications which are targeted for these needs and tied contextually together with other such small apps to produce a cohesive whole.

    My argument is that we need to break this X% windows users market down into a slew of very specific market segments and custom tailor GNOME to be the desktop for the workflow dominant in these specific markets. GNOME has already conquered the generic pc desktop space-it is already *good enough*. Take it to the next step and make it compelling for people in specific occupations to identify themselves with GNOME by giving them exactly what they need. GNOME identities is the key. Simply targetting GNOME at the “enterprise” is never going to produce GNOME identities. Perhaps Pie-in-the-sky, but the only way I can see GNOME actually becoming what people want instead of a not-to-terribly convincing “alternative”. Right now the modern GNOME desktop offers a good “alternative” for the most generic of desktop applications-but nothing compelling for those not already in love with Free Software.

  31. @Mats: I understand Novell, Redhat etc… are targeting enterprises because it’s easier to sell support to a company than to a home user.
    But a lot of the people (maybe the majority) using windows at home don’t work in an environment that requires a computer.
    I don’t even think that if the boss decide to move to Linux all the employers will do the same at home and become open source evangelists. Home users are a much different and wider audience with different needs.

    My first job was in the offices of a bank and all the computers in the room had OS2 installed. But at home I didn’t switch to OS2, because on windows there was much more software and life was much easier.
    Also in a big company usually the desktops are castrated, employers sometimes can’t even even right click or change the wallpaper. If one day the OS will change on their desktops they probably wouldn’t even notice the difference or care about it.

    If you want home users to switch -> target home users.

  32. If you want to know about ways to improve the market share of linux on the desktop, please read the “giant bug-report”, also known as “linuxhater’s” blog (linuxhaters.blogspot.com)

  33. Like others said, Linux need to invest a lot in being admin-free. Having to spend time to tinker on config, or salvaging a broken app, or app updates that eat their previous settings, is absolutely rebiditory be it for business or home users. In market-speak, it makes the TCO too high.

    In enterprise server context app tinkering cost is somewhat offset by rock-solid lifetime with zero admin, but desktop linux does not gets rid of the tinkering while removing all the robustness sofwtare like bind, postfix or apache has

  34. I just recently switched my laptop to Ubuntu when my Vista system was corrupted by a nasty virus and I got sick of the headaches. I have enjoyed it greatly – as has my wife who has now asked me to install it on her desktop. As I cannot get my lexmark printer running under linux, I will keep running Windows on my desktop so that we can print our files – though it stresses me greatly!

    That said, I disagree with many of the software comments here. Sure, OpenOffice and GIMP have SLIGHTLY less going for them than their pricey counterparts, but the vast majority of users, typical desktop home users, do not even touch those features EVER. I have had no problem porting back and forth usable and high quality documents, spreadsheets, pdfs, and more. Further, for those who complain about DVD, codec, multimedia functionality in Linux – it is better than in Windows by far: less crashes, more frequent playback without needing new codecs, and way fewer security holes.

    What kept me for years from making the switch, and what keeps those I know from doing so, is comfort. Plain and simple. There may be gaps where Windows offers better proprietary software, but those gaps are more than erased by the security and usability of Ubuntu (though I haven’t tried other distros since Mandriva was still Mandrake). I knew this, my windows-using friends know this, but comfort reigns supreme.

    Pre-installed linux systems is, then, a step in the right direction as are live CDs. It will be a long, hard battle, but a fight worth fighting. Patience and continued usability improvements are key – it is the usability that has finally brought me back from the dark side.

  35. @Marco:


    But a lot of the people (maybe the majority) using windows at home don’t work in an environment that requires a computer.

    I believe most (more than 50% of) people today needs to use a computer at work, if not daily.


    I don’t even think that if the boss decide to move to Linux all the employers will do the same at home and become open source evangelists. Home users are a much different and wider audience with different needs.

    I don’t recall reading a requirement for using Linux is being an ‘open source evangelist’, and that shouldn’t be a requirement if we are aiming for mass-adoption, anyway.

    All employers will not switch. People(1) like, e.g, you and me care about what web browser, office suite and operating system we are running, but most people(2) just use their computer as a tool.

    People(2) will therefore follow what they use at work, to avoid learning twice.

    When people run Linux at work, the switch will be much easier at home, anyway – it won’t be as unfamiliar.

    Then, after enterprise adoption, perhaps the increased adoption will yield developers, also for the user desktop?

  36. “So if people care about games I think they probably got themselves a Wii, PS3 or Xbox360 to satisfy that need. So I can’t see lack of game support as being the tipping point either.”

    Quite simply, you’re totally wrong here. The PC game market is NOT dying off. It hasn’t even taken a particularly huge hit. It’s going quite strong.

    I still cannot myself name even a single person I know — other than myself — who does not use his PC for gaming. Even my parents both use their PCs for gaming, whether it be for oddball adventure games (my mother) or PC-only FPS titles (my father), or for MMOs (every single one of my friends), or for a smattering of old and new RPGs (again, every single one of my friends, plus my sister), or so on, pretty much everyone uses their PC for games.

    Many of these people do not own a Wii or Xbox or the like, either because those kinds of games do not appeal to them (console games are still very small in scope compared to most PC games) or because they see no reason to spend $600 on an Xbox after spending $1000 on a PC when the PC can play most or all the games they want anyway, and they need the PC for school or work.

    Passing up games is the single stupidest thing a desktop OS can do. It doesn’t matter if it’s only a single game that someone wants to play — if that game doesn’t run on Linux, that’s far more than enough reason not to switch.

    The no cost doesn’t matter to people. Windows came with their PCs, they didn’t even get a choice to not pay for it, so it is essentially “free” for them anyway. Open Source doesn’t matter, because even most hobbyist developers have never actually dug through the code of anything running in their Linux systems. You can’t claim to win on stability — my Linux desktop has more bugs and suffers more random crashes in Firefox, Evince, Rhythmbox, etc. than any Windows XP or Vista machine I’ve ever used — the only issues I’ve ever seen on this laptop are from hibernation, for example, and I already know that Linux is even worse in that scenario on this hardware. You can’t win by features, because the few things that Linux does that Windows doesn’t just don’t add up to a compelling enough reason to switch. You can’t win with hardware support, because on the desktop all those random architectures and weird setups Linux supports are irrelevant and Linux has atrocious support for a good deal of desktop-oriented hardware, including video cards, audio chipsets, wireless cards, printers, scanners, cameras, and so on… Linux supports a lot of them, but all of them, and in the case of wireless nics and video cards, Linux fails to support the most popular items with any satisfactory level. (Broadcom chips appear to be the most common wireless NIC in my shopping research, and Linux has no ready-to-go complete support for any modern video card other than Intels… and even their the Linux support is lagging far behind Windows’.)

    Linux has nothing to compete on on the desktop. Nothing. Not one single compelling advantage to the common consumer. What it does offer is broken or missing hardware support and the instant loss of a wide variety of applications — mostly games — that work just perfectly fine thank you on consumers’ existing Windows machines.

    I’ve setup Linux desktops for a number of people who wanted cheap home-built machines. They have all eventually switched TO Windows — after paying $100 – $200 for a license — because Linux DOES NOT WORK for them. One of the saddest examples is a friend who only wanted three things: a browser, an email client, and Neverwinter Nights. The game NWN even has a native Linux version. But you know, he couldn’t figure out how to install it. Since installing required opening up a terminal, typing in some commands that didn’t work, figuring that out, copy the install script to the hard-disk, modifying the script code, and then re-running it. And then installing the patches required more shell-script-fu, and the user-made content downloads quite often weren’t Linux compatibile. So he paid $150 for a Windows XP Pro license just so he could run the “Linux compatible” game on a system that actually worked.

    Fix hardware support. Get the Intel driver up to speed with the Windows version, and get ATI/AMD hardware fully supported with 3D support at something approaching actually useful speeds. Maybe even get NVIDIA hardware working without the binary driver through nouveau. Get Broadcom to finally release permission to distribute their firmware so the broadcom Linux driver just works. Get hibernation/suspend and power saving features working across the board so laptop users can actually trust Linux on common hardware to work.

    Get Wine fully integrated. Get over the icky feeling it gives it. Make it so that WoW and Half-Life and Oblivion and so on can be installed by just sticking the CD into a freshly installed system and double-clicking the INSTALL icon that pops up in the CD auto-run dialog that should show up like it does on Windows.

    Get that, and then people will stop having many, many solid reasons to switch OFF of Linux. You can’t expect common consumers to want to switch TO it when there are people who’d rather pay big money to get away from it.

    Once you can actually keep the users you’ve already got, then you can worry about attracting new users. Offering better features, better performance, cleaner interfaces, etc. You need to offer more and better without having to put a big “WARNING: you’ll get a couple niceties in exchange for giving up a ton of working hardware, games, and miscellaneous applications.”

    Look at Vista. People hate it because it breaks compatibility with a small number of existing applications. Most consumers don’t give a shit about the DRM or anything. It’s all about how a lot of hardware didn’t have Vista drivers at first and how a number of apps don’t work in Vista’s security model. That’s it. People aren’t willing to give up applications.

    People don’t care about the OS. The OS is irrelevant. The sole purpose of an OS is to run applications. An OS that doesn’t run the applications you want to run is completely and utterly useless, even if it has a wide number of technical advantages over the competition. Who cares if an OS can theoretically run an application twice as fast when in reality it can’t run it at all?

    The Linux desktop people are so out of touch with the regular consumers it’s not even funny. Honestly, how many of your friends are _actually_ regular consumers? Most Linux nerds hang out with other Linux nerds, or convince their friends and family to become Linux nerds. You go through great efforts to get Linux installed for them, get a ton of obscure and complex configurations set up for their specific needs (that Windows doesn’t need, because it Just Works), and then spend a great deal of time answering questions and helping them fix the things that constantly break or don’t do what is needed.

    Or look at the posters above who claims things like “people use what they use at work.” Um, hello? Most people these days get their first PCs as teens (or even younger), and use them exclusively for gaming, Facebook and Youtube (which also doesn’t work on Linux without installing difficult-to-install proprietary add-ons or half-functioning and buggy beta-quality plugins), and for light school work. The people whose first computer experience is at work are last generation. Several generations old, actually. They don’t matter. You are NOT going to get a mass consumer exodus trying to target those people.

    You need to target the people who are buying new computers for the first time (teens), people who are trying to fit in with the “hip” crowd (the Mac users), and people who are generally interested in computers themselves (certainly not the people who buy whatever the office uses).

    Linux desktop developers and evangelists are out of touch. You aren’t regular consumers. You likely don’t even know many regular consumers. Your opinions and beliefs do not reflect what regular people want out of their computers.

    Until you can start realizing that — and start working to understand real regular people — you’re just going to keep pushing goals like “Exchange compatibility” or “100% Microsoft Office compatibility” or a ton of other crap that NOBODY REALLY CARES ABOUT.

    Seriously, the vast majority of office users don’t use even a tenth of Microsoft Office’s features. They don’t. They don’t know how, they don’t care to learn, they probably couldn’t learn even if you tried. I’ve worked support in large “enterprise” environments, and that’s just the truth. People don’t care about Exchange support except for a small handful of corporate environments that aren’t going to use Linux anyway because it isn’t what the big money sales guys are pitching to their computer-illiterate VPs.

    You aren’t going to get consumer by targetting people who don’t care and who wouldn’t switch any way. You can’t target the Enterprise desktop because the Enterprise is full of older folks who started out being businessny way before computers came along and who don’t care to understand the finer points of Linux vs Windows.

    Target the younger generation of home users. In 10 years time, they’ll BE the VPs and CTOs and so on in the corporate workplace. They’ll pull Linux in because it’s the technology they’re already familiar with and already a fan of, because it’s what they’ve been using since they were 12.

    Twelve year olds don’t give a shit about Exchange. When they’re all grown up and pushing technology decisions in their companies, we’d like to hope that they STILL don’t care about Exchange because they’ll be running Open Source mail and groupware servers using CalDAV and IMAP and SMTP instead of MAPI.

    You can’t get people to switch with any real consistency. People don’t like change. People don’t want to give up what they already have. Even if (not when) Linux finally can run everything Windows can, people will still avoid switching because they just don’t need to. But if you get people to START with Linux — and if it stays compelling enough to keep them from switching to the currently more capable (in terms of application support) Windows or Mac platforms — then that same inertia to change will work in your favor. People need a strong reason to switch, and there’s just nothing that Linux (or any OS, for that matter) is ever going to do that Windows doesn’t. You can’t expect people to switch in droves for no reason (and Free Software is not a reason that regular people care about in the least — get over it).

    Quit targetting the old folks and the Enterprise sticks-in-the-mud and target people who are open to new things and open to “cool new tech” and then let the youth of today drive the consumer landscape of tomorrow.

    Or sit there and keep working on Exchange support wondering why nobody is switching despite how you’ve got a list of bullet points two pages long on why “Linux is better” and why all those reasons to stick with Windows (like game support) are imaginary according to all your Linux nerd friends. That’s what you’ve been doing for the last 10 years, so clearly if you keep doing that, it’ll start to magically work some day, right?

  37. The recipe for success is what Microsoft does when
    they capture a market that is held by an incubent:

    * stick in there, make sure you can keep playing for a LONG time
    (at least the FSF gets this part), maybe wait for the market to
    mature/shift/change, wait for that one opening, don’t give up

    * match all of the important features that the competitor has, and
    offer an excellent compatibility story for migrating users
    (MS Excel runs Lotus 1-2-3 macros and IE had a help menu item “For Netscape users” etc)

    * try to build up one or a few scenarios that the competitor does not have (killer apps)
    I think ubuntu is doing great here (wireless switching and 3G is much easier in ubuntu).

    The long story is told above of course. I think it comes down to:

    * Flash needs to just work and not crash (honestly, how hard can it be?
    libflashsupport is open source and adobe IS cooperating)

    * Find a market that really _needs_ GNU/Linux and use that market to
    evolve and grow it to the point that it can break into the harder established markets.

    * Listen to the users: linuxhaters.blogspot.com

    * Fully integrated WINE and MONO.
    * Make windows users feel at home (backspace==BACK in FF)
    * Make sure SAMBA works out of the box (the stuff that ships
    today in ubuntu is not nearly good enough, maybe it would help
    to add better error messages like “cant connect to windows machine
    because of BLAH and maybe then some devs will actually fix it.
    Ironically SAMBA works perfectly between *nix machines but it
    seems nobody is testing the user experience for copying files
    between windows and ubuntu machines)

    * “Twelve year olds don’t give a shit about Exchange”.
    I agree that targeting young people is a good idea. They are not afraid to
    try stuff, they learn quickly and they don’t have tons of data to backup
    before migrating.

    * Higher quality graphics drivers (high performance, open source)
    WHICH ENABLES: new super nice themes/effects and consistency (nice smooth effects)

    * Stop making de-stabilizing features (pulse audio in hardy sucked big time).
    I like pulse but man try to land things smoother!

    * When booting on unsupported hardware or without the correct drivers,
    linux needs to revert to offer some sort of failsafe operation mode.
    This goes for graphics cards and disk controllers etc.

    * Content creation needs to be as strong on linux as content consumption is,
    this includes making sure that it’s trivial and “just works” to record and
    upload youtube videos etc.

    * Someone said “professional rich client apps” but I honestly don’t think those
    are that important. How many CADing architects is there about there?

    * Someone above wrote this (which I thought was an absolutely great comment):
    Going from “less than the sum of its parts” to “more than the sum of its parts.”
    The little touches help fuse the cohesive experience into a synergistically effective whole.

  38. When it comes to attracting casual users such as my fellow college students, we need to take a close look at usage patterns. Most students, for instance, do very little other than basic word processing, instant messaging, and web browsing, specifically YouTube. In my evaluation, we’re doing well on each of those counts besides than the last.

    While I agree with the sentiment, I must disagree with the statement about games. Most families share a computer, and there is at least one person there who wants to play a game. They don’t play the game often enough to warrant buying it. The huge pirate market in software and games means that Windows effectively has more software available for free.

    This is just what I have seen from experience about the same age group.

  39. I actually think all of this is over-analyzing the problem. I think it comes down to:

    * People will never change if they don’t have to.
    * People will change if they see something that compels them. Something that makes them feel like they’re missing out on something. Havoc is right, you have to know what motivates people before you can target them and drive them to want to change.

    People switch from Windows to Mac because they become convinced that “things just work”, or “I can play w/ my photos and home movies faster”, or iTunes will work better with my iPod, or somesuch. And frankly, people stay on Mac at a higher rate because the high cost associated with the switch ends up reinforcing/justifying it in many people’s mind – they will repeat to themselves over and over again how important it is for them to stay on Mac now because they just spent a lot of money.

    Some of us have switched to Linux because of reasons like:

    * We care about our own freedoms above most other things.
    * We want to learn more about the environment we’re in.
    * We value flexibility of configuration over other things. (There’s got to be a way to make these 5 things work together, despite the fact that no one else has ever done it…)
    * We can’t afford the alternatives.
    * We value lack-of-long-term-change above other things. (This is an interesting one because open source projects tend to be “category-killers”, you’re far more likely to use GIMP or Firefox or some other tool in pretty much the exact same way for the rest of your life. You are less likely to have to change what you know. This simply reinforces the notion that people don’t want to change if they don’t have to. However, most people don’t carry enough of a long-term view to actually consider this one.)

    These are great reasons, but they aren’t good enough reasons for most people to change.

    That’s why the place to get it right first is with the New. Netbooks are a great example, but we’re going to lose that battle quickly if more isn’t done soon. Mobile phones are a good example. The more you get people to adopt the New, the more they’re re-think the Old in terms of the New. E.g. you want people to say things like “My old PC just doesn’t work right w/ my hot new Android phone or my cool new netbook – what do I need to change to make it work better?”

    I read a long time ago that Stu Feldman of Bell Labs & IBM fame, the original author of the original “make” utility, had a 5-point system for determining when something was mature. It went roughly like this:

    1. You have a good idea.
    2. You can actually make it work.
    3. You can convince a friend to try it.
    4. People no longer ask you why you’re doing this.
    5. Others get asked why they’re not doing it.

    Given that 5 point scale, stop and think about Linux on the desktop. Where do you think we’re at?

  40. Interesting comments above, none are flat-out wrong (we’re talking about hundreds of millions of people!)

    A few untouched areas:

    There’s a huge opportunity when someone is upgrading their OS, where the choice is: do I upgrade to Vista or upgrade to a Linux distro? The latter is very appealing, and there’s real money at stake. Get people to try it with a Live CD.

    There’s a medium opportunity when someone is migrating to a new computer (which is just about everyone over 12 years old). Yes the new computer comes with latest Windows “for free”. But there’s still real money and hassle at stake for all the old programs that no longer run. Maybe distros could ship with an easy way to virtualize your old copy of Windows 98 or 2000.

    A free desktop has subtle but significant benefits besides price and ideals that aren’t emphasized enough. A lot result from the collaboration that FOSS makes possible.

    * I have *EIGHT* updater programs constantly running on my Windows PCs (Windows, Acrobat, those jerks Symantec, iTunes, Java, Lenovo ThinkVantage Apps, Google). Plus a dozen other programs checking for their own updates like Firefox. It’s ridiculous! One updater for all my software is a huge win.

    * I have Windows’ control panel to configure my screens, but then Intel or ATI adds its own systray icon for their graphics hardware, then Lenovo adds another one for presentations, and they all fight each other. Likewise in my systray I have Windows’ wireless network icon plus Lenovo’s Connection Manager both trying to decide whether to turn on Wi-Fi. Trying to enhance a closed source O.S. is a major pain for hardware vendors and users.

    Keep the faith!

  41. drivers and a zero-problem install experience followed by a non-command line interface that never requires users to access the command line for anything,ever.

    done.

  42. Pingback: Anand Kumria
  43. skierpage said: “Maybe distros could ship with an easy way to virtualize your old copy of Windows 98 or 2000.”

    that is the greatest idea i’ve heard in the “how to help linux flip the desktop” debate for a loong time!

  44. Christian said: “Right now Linux fails in academia use because of simple things like the lack of SPSS.”

    As an academic, I exclusively use operating systems based on the linux kernel, and never had a serious problem. Furthermore, I understand that SPSS is available for GNU/Linux based systems; I’ve not tried it, because it’s non-free. But there is a really user friendly, tried and tested, usable free version of SPSS, called PSPP. PSPP looks like spss, feels like spss does much of the same functionality as spss but it is GPL’d and doesn’t cost an extraordinary amount of money. Some of the advanced statistical methods are lacking, but anyone who understands what those are for is clever enough to use R, Gretl or another more advanced tool.

  45. Funny that I came across this after having spoken to a friend of mine about this. I run Windows on my laptop because I give presentations as part of my job, and my other laptop (similar model, also issued by work) can’t even reliably use the projectors at work most of the time.

    The latest Ubuntu only suspends and resumes correctly on my wife’s laptop with the linux-backports package that pulls in the Atheros driver from 2.6.28. Upgrading at beta time (she’s my test guinea-pig, you see) caused flash to stop playing audio because of a bug in pulseaudio.

    A friend of mine who does Linux kernel work for a living just switched to a MacBook because he observed that after upgrading to a more recent Ubuntu his battery life dropped from 3 hours to 45 minutes – useless for a transatlantic flight.

    I was chatting with a colleague of mine who asked me the question “Are we losing our idealism?”. I think that in a lot of cases, 10 years into this fight, I’m starting to find that the Windows desktop has actually been good enough. So suddenly where I might before have been willing to spend time doing ACPI dumps, and check the bleeding-edge FD.o git repos to see if I could get a driver combination that Just Works(tm), I’m now inclined to just wait and see if another apt-get upgrade will get my test laptop to where it will work. The nice part is that with the Ubuntu 8.10 release, I finally seem to have a “yes” on one machine – my wife’s laptop. If we can start seeing yeses like this, we’ll at least have the people who are interested ideologically happily running their machines. Until that happens, I can’t imagine how we’ll see gains anywhere else.

Comments are closed.