how to become modular

Avi Bryant avi at dabbledb.com
Fri Jul 6 18:33:27 UTC 2007


On 7/6/07, Craig Latta <craig at netjam.org> wrote:

>      The second principle was that discussion of a shared vision could
> ameliorate the lack of a short-term gain, and even hasten the
> implementation of the vision by attracting volunteers. There was a time
> in the Squeak community, it seemed to me, when we could discuss the
> merits of an idea before the implementation was finished. I found it
> useful, and inspiring. This is why I have been writing progress reports
> for Spoon and asking for feedback.

Aha.  My personal observation has been that this principle does not
hold.  One piece of evidence I have is the various version control
systems I have worked on for Squeak.  The current version of
Monticello arose through a series of very incremental and (in
retrospect) "unnecessary" early versions (including "DVS" before it
was called "Monticello"), but each of which was released as a working
and useful artifact without any prior discussion.

For Monticello 2, on the other hand, we've released plenty of
information, tried to open discussion many times, asked for volunteers
at several points, but never released something that people could
actually use for their daily work.  Result: apart from Damien who
recently got some funding to work on it, we've had no response
whatsoever.
This despite the fact that MC2 is a much better and more ambitious
design than any of the prior versions of MC.

I've seen similar patterns with Seaside versions over the years:
discussions about the future go precisely nowhere.  Ditto experimental
branches for people to play with.  But make a deep change that still
lets people get their work done and nobody blinks.

Avi



More information about the Squeak-dev mailing list