Sorry for being late to this discussion. Have been
travelling for the last few weeks and apparently failing
miserably at keeping up with email and meetings...
We went through this whole discussion when Gemini was
first formed, and as a result of many of the limitations
we ended up with separate subprojects. I agree that it
would be great if the ACL committer lists, IP logs,
separate release plans, release reviews and schedules,
etc. etc. were more flexible in the project, but when it
comes down to it, and you have multiple different release
artifacts for multiple different technologies, and the
list of things that are different between them gets so
high that there is not much in common, you kind of end up
concluding that they really are different projects. In the
case of Gemini Blueprint, I think it's fair to say that
the problem wasn't that Blueprint committers were separate
from the rest of the Gemini committers, but that Blueprint
was basically stagnating and inactive as a project,
resulting in a void of active committers.
On 27/04/2012 6:20 AM, Glyn Normington wrote:
At the last RT PMC call there was discussion
of Gemini's subproject structure. Recently, one of the
subprojects temporarily had no committers and the
thought was expressed on the call that this could
perhaps be avoided in future by amalgamating the Gemini
subproject committers into a single ACL. Presumably this
would require amalgamating the subprojects into one.
I spoke to Mike Keith about this and it seems that
it is necessary to keep the subproject structure so
that individual subprojects can be released
independently of each other (which is essential). For
example, the subproject structure enables IP logs to
be managed and approved separately.
So is the conclusion that we can't amalgamate the
ACLs without losing the ability to release subprojects
independently of each other?
rt-pmc mailing list