Dropping the ownership model
Emmanuel Seyman
emmanuel.seyman at club-internet.fr
Tue Nov 22 21:58:27 UTC 2011
* "Jóhann B. Guðmundsson" [22/11/2011 19:28] :
>
> What do people see as pros and cons continuing to use the current
> package ownership model?
Pro: it enforces responsibility. The way things currently work, it's
relatively obvious whose work it is to fix a given bug.
Con: the thing that kicked off the cleanup thread. A package which has an
maintainer that doesn't know how to fix it, no co-maintainers and that isn't
high-profile enough to grab the attention of a proven-packager will not be
updated/fixed/improved.
> Would it be viable to move to something like language SIG based
> ownership of packages?
For every single package? Probably not.
FTR, How is this different from co-maintership?
> As in lower the barrier of entry of contributor without the need and or
> introduction of an package or any sponsorship and have them assigned to
> relevant SIG based on language they either know or want to learn. ( not
> necessarly having to tie packaging with code contribution ).
While I'm all for having more maintainers and co-maintainers, I'ld worried
that lowering the barrier of entry will lead to packagers who don't have the
skills necessary to fix the problems that their packages will have.
> The governing body of the SIG would in essence be the once that would be
> responsible for the components.
For a number of SIGs, this is already the case. As Kevin has said, the KDE
SIG works like this and there's a high sense of ownership in the Perl SIG.
> Do you know or want to learn python. Join the python SIG etc...
I'm not convinced that joining a language SIG is going to help you learn said
language. Seeing diffs of spec files go by won't get you very far.
Emmanuel
More information about the devel
mailing list