Software/geek symbiosis

Yesterday I was lucky enough to get the chance to attend a talk/workshop on TiddlyWiki, presented by its creator Jeremy Ruston, with quality supporting acts from JP and Doc Searls. Jeremy has just joined BT as Head of Open Source Innovation, which is pretty cool in itself.

One of the comments he made that particularly struck me was to the effect that a piece of software is a living thing, but is parasitic, requiring a ‘geek host’ to live on. Much of the value of software isn’t what it does, but what it can potentially do. Once you separate the software from its geek, it starts to die, because it can no longer adapt to changes in its environment.

I think this is something that we’re losing in the world of corporate IT. In the old world the systems the business used tended to be huge, developed in-house, and come with a whole team of geeks attached. Unfortunately in those days we also had heavyweight waterfall processes which made it hard to react to new opportunities for the software to create business value.

These days the big systems tend to be assembled around so-called COTS (commercial, off-the-shelf) products from external vendors. In the real world, of course, COTS products rarely work ‘off-the-shelf’, so there’s usually still plenty of code to be written. Since this is not considered ‘core business’, it gets farmed out to whichever outsourcing company puts in the cheapest bid.

The systems still have large teams attached, but somehow we lost all the geeks. The software’s new hosts are programme managers, project managers, delivery managers, architects, solution designers and who-knows-what else. Actually interacting with the code is considered almost menial – something that can be done by anyone with the appropriate technical skills, regardless of whether they have any knowledge of the specific application or its business context.

It’s much harder to unlock the potential in a piece of software when the people the business talks to aren’t the ones who understand how the code works. It’s even harder when no-one understands how the code works.

On a completely different note, it’s embarrassing how low the turnout seems to be for events like yesterday’s, and the talk by Martin Fowler and Dan North last year. Maybe people are just busy or don’t hear about them, but I can’t help thinking it’s partly down to the fact that we just don’t have many geeks left. On the plus side, at least it meant that those of us who were there got a decent chance to chat to Jeremy and Doc afterwards.