I liked to read this thread and agree with the consensus.
When looking at uservoice I thought: "isn't it like trac with
voting?", so I give a +1 for Erics find.

I experienced the forking problem at least twice. Renaming duplicated
packages to get them indexed as mylab_originalname destroys sense and
references in launch files and depending packages, which is why I'm
still working 'unindexed'. Also those geniuses that are temporarily
with ROS won't get a commit access, so a automatic fork detection and
integration in the wiki page would easily let everybody know about
those developments. At best no one has to insert manual lines like "My
fork is better at this and that: <link>". Having improvements in
several repos/forks is not as bad as guessing if and where there have
been improvements at all.