Got the 3rd person controller "working" in a demo scene, and have continued to think about how selection will work. Also more and more thinking about writing a bespoke controller, just so i know how it works, completely. The tutorials by Jasper Flick are a great resource and starting point. Also, because they use rigidbodys as their underlaying root object, the work on rigidbody sync i've been doing should be able to be directly applicable to the controllers movement.
Oh i found out that the Inter Illusion translation package also has google-translate built in, in that it will take a bit of text and submit it to google for translating. One of the tricks when i did some auto-translating test before was where/how to do it, without having to pay a bunch for it to work. If we just distribute the "raw" or original chat-text to all the clients, and each client then uses its own bandwidth to look it up, that might be better? For some cases of better, anyway :)