peripatetic axiom: 08/06/2006

These are some items from Agile 2006 which interested me enough that I wrote them down.

Peter Coffee's keynote

it's a wonder that spreadsheets don't have features built in to deal with uncertainty
tooling to improve programmer efficiency makes software so much cheaper to write that much more software is worth writing: and so generates jobs for programmers, not the converse
3M apparently has a target for revenues generated by products less than three years old
the value of a codebase is not dependent on the effort that went into building it
nervousness is the enemy of innovation

Agile development manifests the Kolb learning cycle
if you wanted to cerify a team as Agile, maybe you'd only really need to certify the facilitators of the retrospectives
maybe a retrospective (or similar explicit learning activity) is about deliberately moving from unconcious competence to concious incompetence
...and maybe also has something to do with Heidegger's distinction(pdf) of ready-to-hand from present-at-hand. Something that I've had cause to ponder before in its relation to software.

perhaps counter-intuitively, distributed teams tend to have shorter iterations than colocated ones

You are likely in a difficult spot when people start saying things which can be given an "objective" spin but are really claiming the property on the left for themselves so as to imply the property on the right of you:

have the "customer" provide estimates, as well as the developers: when the two are wildly different there's something to learn
the size of a story as esitmated by the developers is independent of its business value
the stories that happen to be in a given iteration are a random sample from the backlog

using puzzle questions at interview discriminates against candidates who aren't white upper-middle class suburban American males.
some white etc. folks find this idea rather upsetting