Some recollections of XP Day 3

I visited XP Day 3. Here are some recollections:

Martin Fowler gets bored Martin Fowler announced that he’s bored with Agile Methods. He said he’s more interested in good design. He went on to discuss the importance of design and how the participants have to care enough about their project to want to do design. He mentioned that the open source community has a lot to contribute on the topic of evolutionary design and that it was a shame that Eric Raymond didn’t manage to get to Snowbird to represent the open source community when the agile manifesto was created.

Scrum Rachel Davies and Joseph Pelrine gave an overview of Scrum and how it can fit in with XP. It seems that we’ve borrowed a lot from Scrum in the XP community. (Particularly the experience of story prioritisation in XP and the rules for stand-up meetings.) It seems as though the big thing with Scrum is measuring estimated time remaining rather than some mythical “percentage complete”.

DSDM - don’t delay buy today! I watched the chairman of the DSDM consortium try to sell DSDM 4.2 (or “enterprise XP” as he likes to call it.) He kicked-off by stating that as a consultant he’s gotten experience of helping out on XP projects that aren’t working. He said that they didn’t work because they aren’t doing the twelve XP practices. That much I can understand, but I don’t understand why a failing XP project would choose to hire a consultant in some other method to fix it. If you had problems with XP wouldn’t you hire an experienced XP coach to help you straighten things out? After the speaker has finished selling DSDM under the guise of XP I’m certain those projects won’t be be doing the twelve XP practices. I respect DSDM but there’s something fundamentally dishonest about selling a method using the name XP when it’s obvious that it isn’t XP. This was the first XP day session I’ve attended that felt like a timeshare pitch.

Retrospectives A write-up of Tim McKinnon’s timeline retrospective demo can be found on Martin Fowler’s Bliki. Apparently, Norm Kerth’s book on retrospectives has a lot of different types of retrospectives that you can use depending on your circumstances.

Tracer Bullets Workshop I also caught the “Tracer Bullets Reloaded” workshop. In this workshop we broke out into two teams and built something. Then we reflected on what we did and tried to do the same exercise again. After reflecting on what we did and how we did it, we knew what job roles we needed and approached the task with brutally simplistic approach which worked a lot better. The advice we wished we could send back in time to ourselves entering the workshop was: “Collect real data as early as possible”. That means try out the exercise as soon and as simply as possible, then work out how to refine what you did. We should have known better but it was amazing how easily we all got led down the “big design up front” path.

Fear and Loathing in JBuilder There’s a breed of corporate technology pundit that has a ready opinion on everything and having tried lots of things in the past doesn’t want to try anything new. I’ve been trying to persuade one or two of these to try a short session (20-45mins) of test-driven development for years. I had pretty much given up on them. I’m happy to report that one of these people did attend the test-first lab. His reaction was interesting: “No - I’m not going to use that cheap rubbish” (I’m paraphrasing here, but he doesn’t like eclipse because it’s free.) Allegedly, you have to create the class that you want to test before you write the test in the version of JBuilder that he has. This is completely counter to test-driven design. When you go to a workshop to learn something it’s better to follow the instructions than to choose to put up extra barriers. My former Aikido Sensei used to say that it’s very important to maintain a “beginner’s mind” - one that is open to learning. If JBuilder doesn’t let you do the work, maybe the code is telling you something. Perhaps it’s saying “beware of expensive imitations”. The fact that he participated is encouraging but perhaps the attitude is a sign that there’s much more work to be done before they “get it”. I hope they “get it” soon because it’ll reduce the amount of suffering in the world especially that part of the world affected by their recommendations.

The programme for XP Day 3 has lots of useful downloads.