Alex Hudson

Thoughts on Technology, Product, & Strategy

Month: October 2008

Bongo at Expo

So, somewhat little advertised and somewhat exclusive, Linux Live Expo 2008 in London has been taking place. Friday was day two of three; for the first time it’s on a Saturday as well – and we’ll see how that goes tomorrow, I guess. But the first few days have been interesting.

Thursday started relatively slowly, although Bongo was sited next to the awesome OpenNMS guys – who travelled a lot further than I did to get there, and are exceedingly friendly. They have some interesting ideas about coloured shirts for contributors and stuff like that which I’m very keen on shamelessly ripping off, and Tarus is an incredible ambassador for their project (and the other guys are ok too ;).

Not long afterwards, I hooked up with the amazing Michael Meeks who was dishing up hip-hop hand-grabs like they were Hershey’s Kisses, and generally being extremely positive to all in the .Org village. Although Novell as a corporation have not seen fit to bless the Expo with their presence, it’s nice to see their more respected engineers about.

In general with Expo, things have been relatively quiet – and somewhat cold & dark, if truth be told – but we’ve made some interesting contacts with different people and projects, and it does seem like it was very much worth attending. Lance and I will be there on Saturday, and that will be extremely interesting too.

Bad day for software patents in the UK

Today wasn’t a great day for software patents in the UK: the judgement in the appeal of the Symbian patent application GB 0325145.1 has been put up on Bailii, and the news is that the appeal was rejected. That means that Symbian’s patent will now be valid in the UK.

For those unaware of the patent, it’s effectively a patent on dynamic library loaders. It allows people to patch the binary interfaces (ABIs) of libraries in a manner which still allows you to load them quickly.

Symbian have managed to pass this off as “faster, more reliable operation” in the computer using the invention. This is basically nonsense.

In technical terms, the computer is obviously operating neither faster nor more reliably: it’s a different way of attacking the same problem which happens to have different results. The “unreliability” isn’t the computer malfunctioning somehow; it’s a result of the limitations of the original design.

It’s relatively easy to critique the application, but you might ask why no-one came up with this before. Like so many software patents, it’s not a question of invention really: it’s a matter of not having faced the problem being “solved”.

Symbian are in a relatively special position. They license out their operating system for others to distribute and modify. So, they are in a position where they need to cope with other people’s changes in a compatible manner.

No-one else really needs to solve this problem, hence it hadn’t been attacked. If you set a programmer this task – to come up with a way of having library indicies which would cope with third-party modifications – you’d probably find they come up with virtually the exact same thing.

So, a sad day for the UKIPO, who in this case were the good guys. The Judges were misled, I believe, by the case put forward by Symbian, since their patent doesn’t really do what it says on the tin.

However, it’s not the end of the world, in my opinion. I think other people will struggle to make the same argument for their patents. I do worry that this is a slippery slope: there is an error of logic in here which cannot be sustained indefinitely, but I don’t think that there is the mood to travel further in the direction of patenting at this point.