EricssonMobileSystems? invited me over to talk about XP in March 2000, during which I ran the StockholmExtremeHour. -- SteveFreeman
The group was quite small, about 12 people, so I ran one team of 6, with the others milling around, which worked quite well. To break up the session, I talked a bit about XP Scheduling then they did a PlanningGame, then I talked about things like PairProgramming and TestDrivenDevelopment and they ran a couple of iterations. At least they got some exercise ;-)
The problems they face way are beyond our usual scope: millions of lines of Plex an event-based language (Programming Language for Exchanges), downtime of less then 3 minutes a year (including hot-swapping operating systems), system testing is still manual (8 weeks!), development all around the world, communication by passing documents around, and just loading the data into a switch takes 5 hours. Also, Ericsson is a very large company with 100 years of history to carry.
That said, they're very keen to move towards more iterative development because their current lead time is just too long for the mobile phone business. They're currently trying to push towards a daily build, which has obvious implications for things like making sure everyone has the right gear and automating the test cycle. As the talk sponsor pointed out, if developers have to load their own data every day, you can guarantee that they will find a way to reduce the time - which the customers have been asking for for ages.
They're also moving towards cross-functional feature teams for new developments. At present, development is based on module owners and this is proving too inflexible and slow.
So, PreachingToTheChoir? here (I also talked to a group of managers and saw quite a few heads nodding), but they're unlikely to be able to take all of XP. That said, they're moving in the right direction and they could adopt, say, PairProgramming tomorrow if they can handle the social issues.
Running the ExtremeHour was definitely better than just talking about XP. It really gave people a feel for the process. I changed the problem from the Tea machine we used in the LondonExtremeHour to a Bread Maker (local customization). This wasn't quite so successful as it took a while for the stakeholders to generate interesting stories. This might be because of cultural differences, but I also think that the tea machine is particularly good because it's conceptually simple but with lots of scope for small variations. Of course, they wanted their bread maker to be enabled for Wireless internet access.
One of the Stakeholders asked me if they could have another iteration. I said, "Sure, but the Nokia Bread Maker is due for release any day now..."