Python for data driven website?

[Alex Martelli]

Peter Hansen wrote:
        ...
> By the way, "We have a preliminary version ... but it is far
> from done. We want to do this right the first time - we expect
> to spend a year working on it" makes me think that you guys ought
> to spend a brief time investigating Extreme Programming and
> considering whether you might benefit from a more iterative
> approach with more frequent releases to get feedback on the
> direction you are taking, as well as to help focus you on the most
> important tasks so you don't get distracted on low business value
> items for too much of the project.

I think they might be better advised to consider Agile Development
Ecosystems more generally -- they're quite a way from XP's sweet
spot, so other Agile approaches might serve them better.

I recommend both Cockburn's and Highsmith's book on the subject
in AW's new series on Agile development.  Read Highsmith first,
is my advice.  Then, with a 3-people project, some variation of
Crystal Clear might well be optimal.

You do have an excellent point, however:

> No project like this should be worked on for three man-years
> before its first release, at least not in this day and age!

Yep, but getting customers interested enough to be willing
to invest manpower is not a trivial problem.  Getting customers
directly involved is the #1 advantage of Agile approaches and
also the hardest one to make into reality.

Still, "getting everything right the first time" is definitely
the wrong mindset "in this day and age".  Rather, focus on
implementing features of high business value first, and on the
ability to perform major refactoring with confidence (through
simplicity and an *abundance* of tests), are very important.


Alex