Gwern Branwen wrote:
There are advantages to it, though. We gain the possibility of people using gitit on top of a darcs repo of articles and codes - and now they have a distributed bug tracker, and distributed doc revision.
All you have done here is mix together permissions that shouldn't be. Write access to the wiki should not be equivalent to write access to the source tree.
Or maybe they do something like set up a repo with web access, and now everyone can edit the source files and improve the haddocks, and occasionally the maintainer pushes the good documentation fixes to the canonical website.
You hardly need a VCS to make a second wiki with freer access and push changes to the canonical site. What gitit buys you is giving people the ability to make a local copy of gitit and 'darcs send'. This is valuable, but let's not exaggerate the benefits by adding fictional ones. The benefit of being able to pull the entire wiki and send a patch to the list is enough to argue that using darcs as the wiki backend has advantages.
Notice that in my email I didn't say that there weren't advantages (I didn't touch on that issue). I simply said that I've seen other established projects struggle with the same speed issue for years. That suggests that the speed issue doesn't have a straight forward solution.
Daniel. _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
