28th
Mercurial’s real main advantage: merging based on changesets?
With a decent experience in Subversion (albeit in tiny teams), I read this tutorial on Mercurial with great interest. My general take is that the real advantage of Mercurial is its merging tool, ie, one that actually works and the fact that it works is because it is based on changesets not on final states. Plus the ‘UI’ of doing multi-level repositories and merges looks more userfriendly and straightforward than in Subversion (where this multi-level approach can be achieved with personalized branches or other third-party tools).
Still, I feel support for Mercurial in terms of clients and hosting services (within and outside institutions) is not on the same level as for Subversion yet, though it is coming (eg, http://www.shelfcloud.com/blog.html). But the client problem might be fatally linked to the GPL-nature of Mercurial. I have also not seen yet how to best split up and recombine large projects (which in Subversion can be ‘done’ by only checking out parts of a repository).
