I am quite fond of reading Bruce Byfields articles. His latest deals very well with the issue of whether the ability to run Windows native programs on GNU/Linux is still valuable or high priority. Over-all I actually agreed with all his points – and have been making some of them myself for a long time.
In the interests of balanced dialogue however, I do feel that I need to point out that there is still a case to be made in favor of the continued development of projects like wine. Of course the first priority should be to use free software on a free operating system. However, even Richard Stallman concedes that sometimes for some tasks which people need to do if they are to earn their living there is not yet a free software alternative and in those cases he says “it is okay to use the non-free program in the meantime, provided you contribute according to your own talents to the projects meant to replace it”.
I see four major aspects that Bruce overlooked in this post. Which by themselves make a strong case for windows application support, especially through wine.
1) The first factor is software for which no functional equivalent exists yet. True as Bruce points out, GNU/Linux has come into it’s own when it comes to productivity software, but where it sometimes lack is for niche-market tasks. Tasks used by a small subset of the computer users out there – often to do mission critical work. This is an issue because those people are now locked into not only a non-free program, but an entire non-free platform. In such small user-communities – there often aren’t any programmers with a free software leaning who could create an alternative – there just isn’t enough of us (yet) to have somebody in every field.
An alternative I find particularly important is the case of clicker. Clicker is a simplified user-interface with a self-tabbing pointer selection which allows for easy creation of picture-sound mapped menus. It’s major purpose: it turns a computer into a communications device for the severely disabled. It is written in visual basic and only runs on windows – and it is not only non-free but very expensive. For many people out there, waiting until the highlight hits the drink icon and then biting down on a mouse-button simulating device is the only way they can ask their mother for a glass of water. This technology has improved the lives of countless. No functional equivalent exists on GNU/Linux. Firstly because the severely disabled do not include a lot of programmers (and the greater majority are children) – and there just hasn’t been a FOSS programmer with sufficient skill and a disabled relative. Until somebody has put in the time and effort to write a replacement, being able to run this app in some way is the only way these people can get some software freedom at this time.
This is just one example. There are many – and until there are none, we must keep developing wine so that not even the marginalized are left out of what is supposed to be the all-inclusive software freedom revolution.
2) Creation of native applications. Google’s picassa depends on wine – but it doesn’t run through wine. Google made use of one of wine’s best features -the ability to rebuild software to link against the wine libraries instead of the windows api – getting effectively native applications. I believe that over time this ability will become almost more important than wine by itself, as it reduces the difficulty for current windows vendors to port their programs by a massive margin. For many companies that have never been prepared to risk the cost of porting for the market-size – the winelib toolkit will reduce that cost to an ever more worthwhile risk. True, the need for proprietory ports will keep reducing, but the all-important transitional phase has really just begun. GNU/Linux only began to make real inroads onto the desktop in the last 5 years, we’re only halfway through the decade. We need to do whatever we can to allow people to switch – every bit of freedom gained, is a shackle shattered.
3) Research: Wine makes public and commodotized what used to be top-secret withheld information of crucial need to a large section of the population (everybody who uses windows). This knowledge is the foundation for much other research. A good example is ReactOS, an attempt to create a free clone of Windows able to run windows applications and behave/operate exactly like the original. ReactOS’s developers were able to focus on their kernel and their bootloader and their desktop – because wine already had their API – and it’s coming along nicely despite a massive delay after (proven false) accusations of code-theft by you-know-who. I probably won’t use it because I like Linux, but if ReactOS reaches sufficient maturity – it may provide a key towards freedom for many who cannot otherwise obtain it.
This is just one project spawned from the knowledge wine made available, many companies still developing for the windows platform are finding wine a better source of documentation on the API they are coding for than the documentation Microsoft provides – because it comes with source. That alone has significant impact – perhaps not directly relevant for GNU/Linux – but certainly not useless (a phrase that should never be applied to human knowledge endeavours).
4) Marketing: Left for last. Bruce makes a very valid case that most of the applications people believe they need to switch are really more a case of yearning for convenience than of actual need, but this doesn’t change the basic reality of the marketplace. To succeed you have to know how to sell yourself. People do yearn for convenience, and they will give up freedom for it if the price in convenience is too high (this is just as visible in many other things in society: people actually beg their governments to censor art galleries, thus removing the artist’s freedom of speech and their own freedom of thought, just for the convenience of not having to judge things for themselves !). So for software freedom to succeed, sadly in this imperfect world, we need to make the convenience loss (at least initially) as little as possible. Merely being able to say ‘GNU/Linux can run photoshop’ counts for far more than the software is actually worth. It means people can switch, keep working and learn GIMP in their own time. Otherwise the fact that the deadline is tomorrow will always trump the ‘but if you just take the time to learn gimp you can save thousands and gain these freedoms’ card for most people – learning takes time, if you have to be unproductive while you do it, it costs money. Allowing parallel learning actually reduces the barrier to entry. I almost had second thoughts about this one, it actually counts against free software in the short term – but I do believe it counts for it in the long term. It’s kind of like the LGPL – a tactical sacrifice made for the sake of gains that outweigh the loss.
That then, the balance-the-debate points I see for promoting wine and windows-support on GNU/Linux. Like I said in advance, I agree with the points Bruce raises – I just disagree with the conclusion he draws from them. I don’t think Google’s investment in Wine is wasted contribution, but I do agree that an investment in an FSF high-priority project would be even better – both however are important nett-gains for the community.
