Fedora clean up process seems to be seriously broken...

["Jóhann B. Guðmundsson"]

On 11/21/2011 10:36 PM, Kevin Fenzi wrote:
> On Mon, 21 Nov 2011 14:03:43 -0800
> Jesse Keating<jkeating at j2solutions.net>  wrote:
>
>> This has come up nearly every release cycle.  Problem is that nobody
>> can seem to agree on what an appropriate "sign of life" would be, no
>> has made a serious FESCo proposal for a contrived sign of life.
>>
>> I don't think anybody disagrees (well maybe KKoffler) that
>> unmaintained software should be discovered and ejected from the
>> distro, the entirety of the problem lies how to discover (as well as
>> side issues about what to do about maintainers that are active for
>> one package, but completely ignore 3 others, etc…)
>>
>> So if you are serious about wanting this fixed, draft a proposal,
>> figure out who's going to do the coding work, and bring it to FESCo.
> To quote Ajax: +!
>
> I think the current policy is not very ideal either, but haven't had
> time/energy to work out a new one. ;)
>
> My last thought was to come up with a automated/script way to gather
> info from: bugzilla, pkgdb, koji, git, mailing lists, etc and output a
> list of 'likely inactive people'. Then, have a group of humans look at
> the list, and try and contact/ping people. With no reply after a
> timeperiod, orphan their packages.

Hum not so sure that will effectively work at least the cleanup process 
needs have take place before we start the next development cycle atleast 
no later then GA so basically the "performance" review of the maintainer 
would have taken place in F15 for F17 and would take place in F16 for 
F18 etc...

>
> Note that we need to balance here cases like:
>
> * maintainer is very active, just ignoring $leafpackage right now.

Indicator that the maintainer needs comaintainers

> * maintainer is on vacation/sick/etc

Indicator that the maintainer needs comaintainers

> * maintainer needs help, we should try and help them out.


Indicator that the maintainer needs comaintainers if not that, workload 
could be spread out to other community groups ( provenpackager/QA etc )

>
> * maintainer doesn't use our bugzilla as their primary bug zone.

That problem can be solved technically as in be made transparent to 
reports and maintainers ( reporters using our bugzilla but maintainers 
using their relevant upstream one )

> * maintainer maintains a software that has a vast number of bugs and
>    they can't deal with them all.

True but you would actually see that on the activity on the bug report

>
> * maintainer is working on higher priority bug, so ignoring feature
>    requests/etc.

Again that would be seen on the activity on the bug report

Encase we are "short" on maintainers one way to increase that pool would 
be to drop the ownership model essentially making everybody 
provenpackager and allow everbody to play in everybody's pool...

JBG