Happy new year everyone.
A question to Matz and all the Ruby contributors: are there any real  
reasons to use Subversion and not Git for Ruby development?

I don't want to turn this into "centralized VCSs suck, distributed  
VCSs rock" kind of discussion, but seriously: many large open source  
projects moved to DVCS in last year or two (as large as OpenJDK,  
OpenSolaris, MySQL, Zope, Firefox, Perl 5), and some of them (Ruby on  
Rails is the most obvious example) have seen dramatically increased  
number of contributions from the outside of the "core team", because  
with DVCS, experimentation of all sorts is so much easier. Others that  
did not are considering a move and in process of evaluation of  
different options: FreeBSD and (again, I may be totally wrong) Emacs  
(leaning towards Bazaar?).

A number of well known projects in the Ruby space use Git now: from  
Rails and Merb to RSpec to DataMapper to Rubinius to even Rake I  
believe (I may be wrong here). RubySpec uses Git, by the way.

Recently series of patches by Brent Roman reminded me again how  
different the process of evaluating of different forks/experiments/ 
patches with Subversion is. Yeah, download 7 files, apply them in  
order, with diff-mode in Emacs helps. I mean, it is not *really* hard  
but it wasn't a no brainer either. Pulling from a person who came up  
with something I may be interested in using Git (and GitHub obviously)  
is so much easier, that you start feeling the difference once you get  
used to DVCS process.

Of course, there is git-svn and hg-svn and bzr-svn of all sorts, and  
they all work fine, but people seems to use what official repository  
uses. Probably because they don't want to bother converting their  
patches, or maybe because converters like git-svn look fragile to  
them, it does not matter much.

What do you think? Is there a way for community to help with this  
transition, if you decide it makes sense?

MK