A few folks have been emailing or IM’ing me about the recent fuss around the Birmingham Library rollout (or lack thereof). I was in contact with the team at the beginning of the project and did help out a little with some training and other bits and bobs but I’m not in a position to comment directly about the project, other than to point out a couple of things that annoyed me about the press coverage.
Firstly all the stories are classic Vendor sports and war metaphors (as Doc would say). Unfortunately they always end up in apples and oranges comparisons. The library project isn’t just a choice of migrating to Linux on one hand and upgrading to XP on the other – the migration is Linux + OOo + Firefox + Thunderbird/Evolution (not sure what they opted for in the end) + … verses a operating system upgrade on the other hand (in fact I doubt it’s even that simple). All other things being equal and without knowing anything about the products involved it would seem likely that changing 5 or more things might cost more (regardless of acquisition costs) than an upgrade of a single component.
Secondly all the articles tend to have the inbuilt assumption that these are simple contained products that have no interdependency and are completely interchangeably equivalent – like the council is deciding whether to buy Bic or Pentel ballpoints. It would seem obvious to buy whichever one is cheaper. Let’s assume for a moment that the Council’s analysis is perfect and the Linux migration is more expensive than an XP upgrade – it’s still justifiable if what you’re getting is better by whatever metrics happen to be important; security, maintainability, usability, long term costs, viability, etc.
Anyway the only complaint I feel comfortable making about the project is that they (AFAIK) didn’t seem to have someone capable of or give permission for someone to engage with the relevant communities, or to put it bluntly they had no-one blogging about the project.
Stephen recently had a brilliant piece called Biggest Community Wins wherein he admits to a slight open-source bias but says the biggest factor in his choice for Redmonk IT is the size of the community around a product / project. I think there is a flip side corollary to that; success in large / complex open-source rollouts is dependent on getting directly involved in the relevant community / communities and telling the world what you’re doing. Something Stephen’s colleague James would call Declarative Living
We have seen that for many business a mental adjustment they have to make is getting used to the code being open – which sounds obvious until you suddenly realise you have to go on stage with no clothes on. You have to admit your mistakes from day one. This is one of the factors that makes successful open source projects successful – the all round transparency helps curb the bulls**t.
Exactly the same level of transparency would help all government projects (IT or otherwise). Help cut out the committee-ism and in the particular case of open-source roll-outs help engage directly with the community whose work you are using. It was in one of the Semasiology of Open Source trilogy (the first one I think) that r0ml made the point that the production of source code is only one part of the cost of enterprise IT – that more time and money was spent in requirements gathering and analysis. All this goes on in these type of projects but (AFAICT) it’s never shared, either on the fly via blogs, mailing list, bug reports, etc, or after the fact in some other form. Not just for the next poor project manager who has to do something similar elsewhere and can learn from your mistakes but because you never know; many eyeballs might make all project management bugs shallow.