TracerBullets
From SPA Wiki
This write up relates to this session --> 1 run at Ot2001 by PaulSimmons and TomAyerst.
Thanks to all who attended. Tom and I got some great feedback at the end of the session. This page is a write-up of the OT2001 session on 3rd April 2001. Feel free to add anonymous comments at the bottom of this page, especially if you attended.
Each team were asked to highlight one issue which they feel relates TracerBullets to their work. In team order, these are:
Its worth spending time finding out where the problem is.
Its too easy to generate a model not relevant to the problem.
Its important to have a stab to get some idea of the problem and scale.
Don't get tool-fixated, don't worry about tools until you need them.
Prod with a stick. Don't get hung up on one hypothesis. Test your hypothesis as soon as possible.
When debugging, find some way to get a lot of debug information for minimum effort.
We're really pleased with this range of issues. Its amazing all this comes from something as simple as probing in a bin with some sticks. Cheers,
PaulSimmons and TomAyerst XPDevelopers and DrKW employees.
Your comments:
8 mins per 40 mins access to the box was about right. Could try something different for practical 2, rather than the same exercise with different object. Its difficult to change your hypothesis of what the object is, during the practicals. This is a reflection on western thinking, rather than a criticism of the workshop itself.
Reflections on issues and comments:
When using feedback to elaborate and use your model look for tests that will invalidate it rather than just keep adding more confirmations that it works. This will push you into new areas faster.
Discussion: (i'll take the liberty to put it just here) I joined the session, had great fun and - as always - questions popped up a few days later.... It seems to me (again after a few days processing) that the session focuses on problem analysis. Just look at the issues raised by the teams. All have problem solving statements in them. It does not easily trace back to the act of developing software and finding the 'right' metrics in that process. It merely states that metrics and small steps are important. There is a gap which i cannot define exactly. -- RobWestgeest
Hmmm, thanks Rob I think your observation is spot-on. I don't think we wanted to discover the 'right' metrics, moreover we wanted to highlight the need to build any cognitive model - even if its wrong at the beginning. -- PaulS
Paul is correct. You need to get 'into the process' as soon as possible. So the earlier we get feedback and start constructing a cognitive model the sooner we can start refining it. Its true that we didn't find any 'real' metrics but there may not be any useful metrics that are universally applicable. Once we get that model going it should, hopefully, suggest appropriate metrics. --TomAyerst
I found this to be one of the most useful sessions in the conference, mainly because it managed to highlight the problems we have with the over-analytical mode of western thought. One of these problems shows up as an issue of tool fixation if you'll pardon the phrase. A conclusion that our small working group reached during the workshop was that we only needed simple tools and better thinking, not better tools with the same old thinking. To me, the issue of metrics can fall prey to this fixation. Good software development is as much, if not more so, about the capabilities of the people doing it, rather than the technology. The gap that RobW mentions may therefore be the gap between the necessary and current 'thinking' capacities of the people involved. Metrics dont help here because we are talking about the development of better human thinking (problem solving) capacities. What this workshop delivered was a forum for the individual to focus on the inadequacies of their own problem solving process. Thanx again. -- CharlesTolman
I just wanted to say that this was a really well-planned and run workshop. It had a simple purpose and a clear way of achieving it. I wish I could have stayed for the whole thing! Well done, guys. -- JohnDaniels
