"Agile Architecture" at Agile 2006

Last week I was at the Agile 2006 conference, in Minneapolis and I see from the records that some of you were too: hello again to anyone I met there.

My session was in the first slot after the opening keynote on Monday morning, and I got the impression that some of the folks there hadn't fully grasped what the nature of a Discovery Session was, some other presenters found the same, I think. I believe that the origanisers might do more to help attendees understand better what's likely to be expected of them at such sessions for next year: they aren't tutorials. Of course, my session was a re-presentation of an XP Day session, which like Spa sessions (the other sort I'm used to presenting) tend to be exploratory, open-ended and interactive in a way that the typical Agile session seems not to be.

Outputs

As soon as I get unpacked from moving house I'll have the London reuslts here to compare with, but a selection of the Minneapolis results are here.

The session was bookended by very quick brainstorm to try an capture any change in the attendees thinking after the discussions and work with the lego.

Before

TDD-not fit for mainframe volatile develop platform framework shared direction drives communication developers like complexity correct today testable malleable adaptable think big, start small

standard organic correct today testable design isn't dead! nimble easy to change business engagement high test coverage flexible easy to change dangerous

And a couple of longer items:

• Have an idea of what possibilities your customers may need to explore
• Not for conventional projects, developers know too much business and not only code by success criteria
• provide guide rails for teams to make decisions within

Indeed, developers do like complexity. Given half a chance, they will put wings on a train!

After

experiment results the right solution for the right problem do only what's needed to satisfy requirements testable not holy ground willingness to restart different is OK ROI

cost of implementation time to implement brilliant think big, start small responsive correct today (still) locally optimal flexible (within reason) mutant

It looks to me as if the aims of the session were met.

Some conference wiki also bears some comments on the session.