NS Internet column
Written 11 October 1999 for the New Statesman
A friend of mine is writing a book about how Linux and open source software generally, is going to take over the world; but he refuses to say which word processor he's writing it with and this is of course because he's writing it under Windows. It reminds me of the first time I was immersed in the world of crypto freaks, and discovered that the people developing PGP 3.0, another program that was going to change the world, did not use it to communicate amongst themselves because it was too much hassle. Given the paper fortunes that are being made with Linux software it is quite interesting that everyone who uses it admits, if pressed, that it is eight or nine years behind Windows in simple desktop application software.
I only noticed this myself by accident, so blinded was I by delight at getting a working Linux system running. Once I had got it doing almost all the things I wanted, I suddenly noticed that I had not booted it up for a month simply because it did none of them as well as Windows does. It never crashes, but this is not much use when I have to shut it down even to write a headed letter or to browse the web without crashing. All the pleasure I had had from it came from getting the damn thing to work at all; none came from getting work done without fuss.
This will change, I think. In a year or two there ought to be Linux programs that do most of the things that Windows programs now do, and which will be designed so thoughtfully that they won't need manuals. (This last is important because they won't get decent manuals). But for the moment the uselessness of Linux for most practical purposes is one of those things that everyone knows but no one can admit. I have been playing around dropping remarks like this into various Linux discussion groups, partly for fun, and partly to discover if I had missed some category of wonderful programs that would make sense of all the hype. I haven't. "It's a hacker's operating system, for hackers" says one defender. "It's about where Windows 3.0 was in 1990."
Both these remarks seem to me indubitably true: why, then, are ordinary users expected to go back eight or nine years in their computing experience? Part of the answer is a rabid hatred of Microsoft which is not to be believed until experienced. It's rather like the way the American Right hates Clinton in that its object is neither loveable nor perfect, but could not possibly deserve the vituperation heaped on it. Sufferers will not admit that anything Microsoft makes could be superior to any of the competition. I'm not a fan of Microsoft myself, but the fact remains that some of their software is excellent, and their behaviour, however reprehensible it often is, seems right in the mainstream of accepted American business practices.
More important, I think, is a reason that goes to the heart of the weaknesses, as well as the strengths of Open-source software, like Linux. The really successful open source projects have been built and maintained by their users. That is one reason why they are so common in intrinsically complicated fields like networking. If you can understand what you are doing with a network you probably need to understand what the programs that help you do it are trying to achieve and, if they fail, try to fix them.
The situation could not be more different with desktop software. The people who actually use a web browser need have no idea how it works or what it does: that is the single most important reason that the Web replaced or absorbed all other protocols on the Internet so quickly. This fact explains why the greatest Open Source fiasco to date has been the attempt to replace Netscape Navigator with a version which was not merely free in the sense that it was given away, but also free in the sense that the source code comes with it, and anyone can play with it as they want to. It has taken eighteen months to produce something which does half what its predecessor managed for an hour or two at a time before collapsing. It could well be another six months before the whole thing is finished.
Similar considerations apply to word processors, spreadsheets, and so forth. A man capable of writing a word processor no more knows what makes a good one than I know what makes a good programming environment. In a market, I have a way of making my wishes important to him; under the Open Source system, the approval of users is worth far less than the approbation of his peers, for whom "user" is just an abbreviation of "luser".