Agreed with all you write.
I've been looking into several projects and programs doing what we do here, i.e. the "translation" part of localizing software. Only OmegaT works with glossaries, but I do not know about any free glossaries related to the kind of terminology we need, so that needs research.
We must perspectively look into building or own glossaries. Good is that they can be very specific. I doubt :-) references to usage contexts such as "Mediawiki extension 'Socialprofile'" would be accepted in more general wordbook projects. They are however in Ambaradan, and maybe in Omegawiki, should we choose to automatically share collected data.
Let's see, what I am going to find, while doing some glossary related tests and research anyways in the future. I shall report.