I am of the opinion that "umbrella TLP's" that aggregate loosely related subjects are a bad. The most visible such project at Apache is Jakarta, and it is taking years to dissolve it, moving active projects to TLP's and inactive - to the Attic. ASF doesn't have a limit on the number of TLPs and the Board is actively encouraging each distinct community to orm its own.
To me the main criteria for a subproject would be a significant overlap in the committer/PMC base between the 2 projects. Otherwise we are just impeding the community development IMO. Consider that DirectMemory becoming a Cayenne subproject would mean that Cayenne PMC will need to vote on DirectMemory releases. This doesn't make much sense with the current PMC composition (we have no people involved in DirectMemory development on the current PMC). If we vote in DirectMemory folks to the Cayenne PMC, and they will vote on DirectMemory releases, while we'll vote on Cayenne releases, we still don't have a cohesive PMC, and there's no reason why DirectMemory can't be a TLP in that later case anyways. Hope this makes sense. Andrus On Sep 20, 2011, at 5:04 PM, Christian Grobmeier wrote: > On Tue, Sep 20, 2011 at 3:56 PM, Andrus Adamchik <[email protected]> > wrote: >> Looking at my reply, the first paragraph came out a bit brash. Didn't mean >> it to sound that way :) > > no worries, i didn't get it wrong :-) > >> Certainly a great proposal, and looks like it has already generated quite >> some interest in the Incubator. So good luck, and keep us posted. We are >> definitely interested to try it out as a Cayenne cache provider. > > Will keep you posted. Anyway any other Cayenne committer is welcome > to, either as dev, mentor whatever. > > Another question- DirectMemory is currently looking for a sponsor. One > of the thing which has been discussed internally was if Cayenne would > sponsor it. After incubation it would be expected that DM is an > subproject of Cayenne. As the implementation is basically a general > one, I have some feeling it would make a good TLD. On the other hand, > it can be made a TLD later and could start as Cayenne-Subproject. This > might have benefits for all sides. > > How do you (all) feel about this? > > Cheers > Christian > > >> >> Cheers, >> Andrus >> >> >> On Sep 20, 2011, at 12:29 PM, Andrus Adamchik wrote: >>> I can't mentor any more projects. I am mentoring Wave and also incoming >>> OpenMeetings. >>> >>> The idea is great, and as long as it is not as invasive as Terracotta, >>> we'll definitely be interested to try it out and give feedback. The >>> usability of DirectMemory in Cayenne depends on >>> >>> * the API model they provide to the applications >>> * whether we can build a cache adapter with groups support >>> * on speed of serialization with LOCAL and SHARED cache. (I'd imagine it >>> will be more appropriate for SHARED, as it stores DataRows with no >>> relationships and detached from the context). >>> >>> Andrus >>> >>> On Sep 20, 2011, at 11:50 AM, Christian Grobmeier wrote: >>> >>>> Hello folks, >>>> >>>> a new proposal will arrive at the incubator the next few days: >>>> http://wiki.apache.org/incubator/DirectMemoryProposal >>>> >>>> It is a cache implementation similar to OSCache. Now I know about the >>>> discussion of the OSCache usage the past weeks and thought, this might >>>> be a good chance for Cayenne to bring in own needs etc. >>>> >>>> I told the guys that Cayenne might benefit from this new lib. They >>>> liked that idea very much, so it seems. >>>> >>>> Anyway, they asked me if some of you would like to step up to this >>>> proposal, either as an initial committer or mentor. I think this is a >>>> great opportunity! >>>> >>>> Please let me know so I can tell the people and sign up at the >>>> proposal link above. Any questions, shout! :-) >>>> >>>> Cheers, >>>> Christian >>>> >>> >>> >> >> > > > > -- > http://www.grobmeier.de >
