Here is a recent (short) article comparing pgsql and mysql.
http://www.informationweek.com/story/showArticle.jhtml?articleID=2390234
6&tid=5979
Overall, not a bad article, although the claim of missing row locking is
a familiar irritant. It would be nice to get some fair treatment on
this point once in a while.
I've mentioned before that the MVCC treatment in the manual is light and
doesn't bridge the gap between academic principles and user
understanding (any by extension, the technical press). I think it's
time to do something about it, perhaps with a patch to docs, etc. Does
anybody else think this is worthwhile? I'll have a go at it but I'm not
really sure I have the chutzpah to pull it off.
Merlin