Open Stack

Thomas Goirand wrote:
> IMO, we should, for the Liberty release, get rid of:
> - suds & suds-jurko
> - memcached (in the favor of pymemcache)
> - mysqldb (this has been discussed at large already)
> - cliff-tablib and tablib (not ported to Py3, used only for testing)

I feel like there is an opportunity for a project team to form around
generally "keeping up with the Python times". In the past we had various
disconnected efforts: Python 3 migration, dependency convergence,
getting rid (or taking over from) abandoned dependencies.

All those efforts share a lot in common: you have to keep track and be
closely involved with the Python ecosystem at large. You have to see
what is getting abandoned or will never support newer versions of
Python, check out new libraries that could replace things we use,
actively push for those replacements in projects across OpenStack and
generally work to converge on dependencies where possible. You have to
watch upper-constraints and bump to keep up with bugfixes and avoid
falling behind.

Up to now those were mostly individual efforts. It may be time for a
true horizontal effort to form around those activities and skillset:
establishing an official project team would help coordinate those
efforts and give them some weight. The same way Oslo reduced code
duplication, I can see that new effort result in increasing the
packageability (ew) of OpenStack and serve as insurance against falling
behind the times.

Thoughts ?

-- 
Thierry Carrez (ttx)