Solaris 10 shines, but there's a risk of burning

Anyone who still doubts the power of open thinking to change the world need only read our review of Solaris 10. Solaris is a mature, complex operating system that embodies some of the latest thinking on OS design from a company with a long tradition of excellence in computing architecture. These used to be good enough reasons for Sun to insist on big bucks for hardware and software.

Now, Sun is rabidly keen on giving Solaris away to run on commodity hardware: a revolution as significant as Copernicus, but in reverse. From Sun considering itself the centre of its universe, it's now in orbit around Planet User.

This is a tremendous advance. But like any great shift in theological doctrine, it's not without its risks. By making big claims — compatibility with hundreds of x86 systems — that cost nothing more than a few hours to test, Sun may find that people slap the OS on a few spare PCs only to hit device driver problems — as shown by the compatibility issues flagged up by our review. They'll walk away with nothing but bad memories: there's nothing like a nasty burn from a morning's incautious exposure to put you off sun-worshipping for life.

Brilliant operating systems from industry leaders can easily be brought low by niggling problems that poison the user experience — if it won't load, it won't go. Just ask IBM about OS/2. There is only one way to avoid these on a platform as resolutely mutagenic and heterogeneous as x86, and that's to spend a huge and continuing amount of effort in keeping up. Microsoft manages it by sheer corporate sweat, open source by sheer community willpower.

You can't just transplant one into the other: expecting a community to sign up for corporate purposes without giving as much in return will be an exercise in expensive futility unmatched since the height of the Soviet command economy.

Free and open source software works because it is a community investing in itself. To join up — and to reap the rewards — requires a much bigger investment than just slapping the labels on. Nothing would give us greater pleasure than seeing Solaris 10 shine brightly, attracting widespread support and leading a move to efficient, powerful and cost-effective open systems, and Sun's claim of a million downloads to date is a great start. Sun can also point with pride to significant contributions to the open source cause: the world would be very different without OpenOffice.

Yet it isn't churlish to say 'good, but not good enough'. For Solaris 10 to be all that could be wished, Sun must be objective about its deficiencies and swift to find ways to remedy them — and if that means giving the open source community tools and motivation, so be it. Otherwise, the company risks being seen as not quite serious about its commitments to the community; either way, you can be sure that the community will reciprocate.