Fedora Project Board Meeting (2006-05-02)
1. MaxSpevack sent an agenda out for review.
2. Summer of Code applications have begun to filter in. JeremyKatz is going to be a mentor.
SethVidal and RexDieter brought the rpmlint idea to Fedora Extras Steering Committee (FESCO). A summary of the feedback was:
- We need to find a balance between what a computer can do and what people should do. Certain things, like the summary and description, must be reviewed by a person, but some of the packaging guidelines could have automated systems determine compliance.
This led to the next question: how can we get through the backlog of packages that are waiting to get into Extras?
- rpmlint wouldn't necessarily speed this process up.
- There needs to be some interest in the package for it to move through the process.
- People who submit packages need an answer in a timely fashion.
- It was suggested that the package review process doesn't scale, and that we need to push the problems back to the packagers.
However, this raised a few counterpoints from within the Board:
- It was not universally accepted that the current process *doesn't* scale.
- Nor was it universally accepted that a package backlog is even a problem that needs to be dealt with. Is the backlog actually preventing anyone from submitting good packages, or are the best packages filtering to the top anyway?
- We need some metrics to arrive at an informed opinion:
- ElliotLee can help with bugzilla queries (how fast the queue is growing, etc.)
- SethVidal and RexDieter own figuring out whether or not we need to clear this backlog better, and if so, putting together a plan for doing so.
Updates to packaging guidelines was brought up as an example of how technical decision that impact both Core and Extras are made, now that we are actively working to adopt a single set of guidelines for both.
The suggestion of a Technical Steering Committee was raised, with a membership that is completely balanced between Red Hat and non-Red Hat members.
Rather than get bogged down in the logistics of such a thing, we decided it would be better to build up the organization as there was work already happening or needing to happen, and letting it evolve as naturally as possible. Right now, we need a group that can own the pacakging guidelines. Is this group a strict subset of folks who are already a part of FESCO?
- RexDieter will continue talking about this kind of stuff inside of FESCO, and see what people there think.
- MaxSpevack and other RH folks will build internal consensus.
- MaxSpevack and SethVidal need to drive the next iteration of this discussion.
Fedora Core 6
How do we increase transparency into what is happening on a development front? There are numerous reasons why this is a good thing to do, including but not limited to the fact that it increses community involvement, increases the number of ideas that are on the table, and encourages open dealings and decision making.
- MaxSpevack needs to figure out a way to give updates on FC6 without while managing expectations and not making too many people unhappy.
We may have LukeMacken do some infrastructure work this summer, but I don't know if anyone has told him that yet.
- MaxSpevack will put together a job req that can be used to get contractors, interns, etc.
In short, we have a strong need, and not a lot of extra time to look into it. ElliotLee looked at Savane , and it's a potential option, but will have hardware needs and require sysadmin work.
- ElliotLee needs to drive us in the direction of a solution, ideally by leading a group of folks who are interested in making real progress in solving this problem.