What to do when open source is not good enough
Open source is not always the best software. When it's not, use something else. Is that so complicated?
The obvious answer is "Use something else." Some, however, don't like simple answers to obvious questions, so I'm now going to spend 1,000 words or so saying, "Use something else." You have been warned...
For a variety of reasons, I've been thinking lately about what to do when open-source software is not good enough for a given set of needs/requirements. There are some who believe that we should use open-source software, even if it's rubbish. I'm not among that group.
Fortunately, this is less and less an issue as open-source software becomes better--even superior, in a growing number of instances--to proprietary software. I use Adium (open-source instant messaging application) because it's better than Apple's iChat (though I turn to iChat when I want to do video chat because, well, Adium doesn't offer this feature). I use NeoOffice (OpenOffice for the Mac) increasingly because I actually prefer its presentation program to Microsoft's PowerPoint, but I have to head back to PowerPoint when I want to embed video in a presentation.
And so on.
For those who think that open source is something to impose, I disagree. I believe that open source can carry its own arguments. When it can't, I don't use it, but hope for people to come along who will fill the void. They almost always do.
Take the Linux desktop. The Linux desktop is currently not on par with its proprietary cousins. It's not, no matter how much one may pretend otherwise. Guess what: it's not my job to pretend otherwise. It's not even my job to improve it. Red Hat, Novell, Sun, Canonical and others have taken up that role. My job at present is to make open-source content collaboration and management software better than proprietary alternatives, and I (or, rather, my company) am succeeding. In spades.
Until free software, like the Linux desktop, is a net benefit rather than a net drain on my productivity, I don't see any reason to use it. As I wrote back in 2005:
I aggressively promote Linux where it's a no-brainer (data center, edge of the network, embedded devices), and somewhat cautiously promote it on the desktop, where the average user resides. I don't want to have anyone get suckered into using a system that is not yet as easy to use as a Mac or Windows machine. There's no sense in burning that bridge. Linux desktops have a buying audience today (engineering, fixed-purpose desktops), which...will continue to grow as we make it better. Use it where it fits best.
I think it is foolish and counter-productive to lure unwitting people into an open-source experience that will underwhelm. Why not focus on the areas/applications/projects where open source shines, rather than suffocating would-be converts with where it is murky?
So, instead of fetishing the "Linux desktop," why not focus where open source is already strong on the desktop? Namely, applications. Firefox (on Windows, Mac or Linux). OpenOffice (on Windows, Mac or Linux). Adium (on the Mac). GIMP (on Windows, Mac or Linux). Etc. The open-source desktop is already here--we just get so caught up in the underlying operating system that we forget this critical fact.
In similar manner, the more we fixate on "open-source software at all costs," the less we recognize that open-source software is already providing exceptional software throughout the market. We haven't 100 percent covered the market yet, but we will. In the meantime, the best course is to use what works best. Often this will be open source. But not always. (Like Mark Shuttleworth has written about interoperating with proprietary drivers in otherwise free/open Ubuntu, you use the alternatives available to you.)
At those times when open source isn't the best, we do prospective users and, hence, ourselves, a disservice when we try to convince them that they should use substandard software just because it's open source. That's a losing strategy, and one that we shouldn't use. There are plenty of winning arguments in open source. There really is no need to resort to lame ones.