| Debriefing |
It had problems such as an overly
rigid notion of what the entities
were. My thirst for elegance prompted
the move to a more knowledge-based
approach: the Second TIE Prototype.
My current perspective is that I should have
stuck with this code and explored the
features especially unique to TIE: criteria
and evaluations. Jumping ship offended both
the Extreme Programming dictum:
do the simplest thing that could possibly work
and the Linux mantra:
Release early, release often. Neither of
had been articulated in print yet, AFAIK.
Anyway, I was playing the better-is-better game
rather than the worse-is-better game, so that's
the trap, isn't it.
My partners did voice some concerns along these lines. Sigh.
|