[CMake-Promote] community responsiveness

[Brandon Van Every]

Recently I went on some field expeditions into other build tool
communities.  I was looking for advantages that other tools may have,
particularly their choice of scripting language.  I soon noticed
something interesting, however.  Or should I say boring?  :-)  Most of
the build tool communities I went to were a lot less active than the
CMake community.  At least, their mailing lists had a lot less volume.
 Crickets were often chirping, and when someone did post something, it
was usually pretty boring to read.  Seemed like a sort of "hands-off,
don't speak unless there's a bug, we're only talking in order to file
bugs" approach.  Granted, some of the build tools had multiple mailing
lists, for users vs. developers vs. announcements and so forth.  But
even developer traffic seemed sparse, to the extent I bothered to
look.  CMake seems to have a critical mass of interested users and
developers that other build tool communities lack.

Kitware's willingness to interact with the CMake community, to help
people with their problems, to act upon community input, and to take
patches, is a major strategic advantage.  I recognized this awhile ago
when I first joined the CMake community.  It was a big part of why KDE
dumped SCons in favor of CMake.  I'm surprised that more build tool
communities aren't like this, but it seems to be true.

Is there a way we can put the communication value of the CMake
community front and center?  What demonstrations or metrics might sway
people?  What would put them off?  Claims about community dynamics
being "better" need to be stated carefully.  Most techies resist
anything they perceive as a label, and look for ways to refute the
label.  On the other hand, if we can't make a convincing sales pitch,
then people will only discover the community benefits by trying out
multiple communities.  That's a very slow process.  It would be better
if we had a way to nudge people into our camp first.

What if we measured and put on a website:
- the number of first time posts on a new subject
- the time elapsed until someone answers that post
- the number of first time posts that go unanswered

I bet if people compared objective measurements of the CMake
community, with their subjective impressions of other communities,
that the CMake community would look pretty good.

A willingness to measure this might also look good to managerial
types.  Perhaps we'd also be perceived as striving for more
responsiveness, even if we're merely measuring it.


Cheers,
Brandon Van Every