all code is in english(important code at least) therefore it makes sense for developers to go to english until the program is stable enough for the tinkering required.
besides, you are replying in english, what are you complaining about?
Internationalization =/= translation
Another problem with localising it would be the translation of certain terms.
Like, take the term Mountainhomes.
In Dutch that would become 'Bergthuis' which sounds really silly because it's singular(there's no specific multiple form for home in Dutch).
So you'd have to find an alternative. Like... 'Thuisgebergten' (Homemountains/Homemountainrange)
Or 'Berg gehucht'(MountainHamlet) 'Berg burchten'(Mountain Fortresses) 'Berg paleizen'(Mountain Palaces).
And not to even start on things like goblins and kobolds.(In the Dutch translation of Harry Potter the goblins at gringotts were translated as kobolds)
This isn't really Toady's problem, is it? It's the translators job!
Translation isn't transferring the text 1:1, it almost always includes making compromises, using different terms, coming up with your own words. As for Chthonic's question - it depends heavily on the translator's quality. Some simply keep the English words, some come up with their own or find similarities in local folklore. We have a brilliant Czech translation of the Lord of the Rings, for example, and the lady who translated it made up amazing and naturally sounding creature names or Hobbit names. (But the Czech fantasy norm is heavily based on English, yes. We mostly use English creature names, with only slight changes sometimes to make them easier to pronounce. That's because fantasy comes in the form of computer games (people used to play only in English, though translations are lately becoming frequent) and bad books (long live the Warhammer) translated by horrible translators.)
The only problem from the perspective of Toady I can think of is the way names are created, and how names are translated from dwarven/elvish/etc. to english. Basically, it just takes two random dwarven words, puts them together and transfers both to English (so you have surnames like Clasplashes, Keygem). This works in Germanic languages, I suppose, but not eg. in "our" Slavic languages. You can't just mash any two words together, in fact, mashing words together is very rare compared to English. If you did that, it would sound very weird. Not weird as "Clasplashes", but weird as in "totally ungrammatical" or "outright wrong". And I have no idea how to handle this.
To elaborate:
English "Clasp" + "Lashes" = "Clasplashes"
Czech "Mihotat" + "Řasy" (direct translation) = Řasomih (it sounds horribly, but could be taken with a grain of salt because it's "translation from dwarvish")
English "Key" + "Gem" = "Keygem"
Czech "Klíč" + "drahokam" = Klíčodrahokam (easy, but sounds horrible), more probably "Drahoklíč" (sounds strange), most probably "Klíčový drahokam" or "Drahokamový klíč" (making one of the words adjective, which could actually be simple)
The problem you'd have to code the grammar rules for creating names in languages that can't use simple mashing. Which would require a lot of time and a good knowledge of the target language. I know the names are a very minor part and could be ignored (= done badly) but perhaps Toady could somehow export the grammatical rules to raws and let us program our own? It's the only solution I can think of.