On Sat, 17 Feb 2007 00:52:00 +0100 "Diego 'Flameeyes' Pettenò" <[EMAIL PROTECTED]> wrote: | How much good did Google Summer of Code to the Gentoo _community_ ? | How much the project were of use for the Gentoo (Linux) project?
Perhaps it should serve as a lesson when selecting next year's projects. At least half of the accepted proposals were doomed from the start due to their being either far too much work or far too large a change. Of the proposals that were doable, some of them ran into rather unfortunate external snags of the "people ending up in hospital" variety. The rest... Hard to say. So perhaps next time around mentors should accept concrete proposals that are really doable, rather than stuff that just seems shiny. Gentoo *could* benefit greatly from SoC, if things are done properly. How about a list of criteria for things that are considered sane? Here's a first stab at this: * Nothing that involves changing lots of the tree all at once. (Things that require migrating the tree over gradually are known to be doable.) * Nothing that depends upon substantial input from any particular group within Gentoo. * Nothing that's a brand new project. Building upon existing code is fine (heck, even things like eselect modules), but from-scratch "build a framework and use it" type projects should be out. * Nothing that's an all or nothing type affair. If a partial result is of no use, the whole thing is of no use. * Nothing that requires lots of prior learning. There are plenty of useful things that could be done within these guidelines, and there's a good chance that they really would get done. I'll give some examples of things that I know are doable (biased towards things I know, obviously): * devmanual. Not converting it over to a new shiny XML thing or whatever, but just extending and reworking the parts that need it. * eselectification of some of the remaining -update utilities. * Extending existing QA tools. The qualudis extensions could very well have worked last time around. -- Ciaran McCreesh Mail : ciaranm at ciaranm.org Web : http://ciaranm.org/ Paludis, the secure package manager : http://paludis.pioto.org/
signature.asc
Description: PGP signature
