I wrote a tool to make punctuation in gettext translations match the
original text . It includes a simple set of rules that cover most
common languages, but probably not all. I jumped the gun somewhat and
applied it to all languages (except Japanese and Chinese) and it
didn't fare so well with everyone  for obvious reasons.
At any rate, I'm planning to use the tool on the Danish translations
and commit those changes tomorrow. The respective maintainers for
other languages are hereby encouraged to try out the tool and possibly
make changes if it makes sense.
Also, I think it's worthwhile to look through the cases where the
original text is missing punctuation such that this kind of tool may
fare better overall.