Archive for September, 2010

A tale of the textile and software industry

Monday, September 20th, 2010

A couple of weeks back I was on a panel at the Swedish Multicore Day in Stockholm, with the subject “Multicore technology: what’s in it for Sweden?”, chaired by professor Erik Hagersten (also a successful entrepreneur: his company, Acumem, is doing pretty cool stuff). The setup itself was a sign of new times: Microsoft and Intel were represented by two Swedes who don’t live in Sweden (Per Hammarlund and Niklas Gustafsson), while the Swedish industry was represented by myself (a Hungarian, not living in Sweden neither); the only “odd” presence was Christer  Bengtsson from Swedsoft (who is a Swede and does actually live in Sweden).

The discussion that was spun was quite interesting. It was remarked that quite many brilliant Swedish engineers chose to emigrate (why?), that there’s no push for excellence in the Swedish academia, with only a handful of really world class researchers (why are people happy to be the big fish in a small pond?). Myself I stressed the need for a three-party partnership where universities improve dramatically their quality (e.g. by attracting top notch researchers from elsewhere); the industry steps in with resources and problems that, once solved, can create a competitive advantage; public agencies direct their funding towards supporting excellence instead of general, average quality.

What really triggered this post however is the comment a distinguished colleague of mine, Mike Williams (of Erlang fame), made during the debate: he wants us to find ways to keep software development shifting irreversibly to India and China – e.g. by being so efficient that this shift would not make sense.

Is this a fight worth taking?

Obviously, making software development more efficient makes perfect sense and it’s something that is very high on the agenda for most companies. But, can we (Westerners) really win on this front? I doubt so – I don’t think is the way to stop jobs shifting to China or elsewhere. Just take a look at the textile industry: you would be hard pressed to find cloths manufactured in Sweden or Finland: these countries simply cannot push the cost low enough to justify production locally (it’s not even justified anymore in countries like Moldova or Ukraine neither). I strongly believe the same will happen with mainstream, bulk software development – and it’s really the way it should be, at least from an economics perspective. It may hurt one’s national feelings, but it will not be the end of the Western civilization: new opportunities will emerge and the society will adjust, as it did when the textile industry shifted elsewhere.

So, is this the end of the software industry in Europe?

Let me answer with a question: did textile production stop completely? No. High (very high) end cloth designers did not move altogether elsewhere; they still have their shops in Milan, London or Paris and they make good money. The whole industry shrank but moved up dramatically the value chain and the social environment of Europe (and certain parts of the USA like Silicon Valley) – quality of living, environment etc – made sure that it’s the place of choice for many of the brightest minds. I strongly believe the same will happen with the software industry: in 20 years, there will be far less software designers in Europe, but they will work on top notch projects (space, avionics, all sort of mission critical applications etc) where individual brilliance matters before anything else – and those brilliant software designers will choose to stay for the same reasons great cloth designers are not working elsewhere. As for the rest, it will follow the natural laws of economics and shift to where it’s most efficient to produce. It would be a mistake to fight this trend – it will just happen; instead we’ll need to focus on those areas where brilliance rather than cost matters.

And make sure our social environment is still a top attraction for the best.