Arthur,
If you'd like to make changes to the data that's included in the format I'm more than happy to discuss it; the only people using that information must be you and me anyway ;-)
Well, what I was wondering is if we need to load any data at all (except english for compatibility). If it can be loaded at runtime, then this is the best place, because the format itself doesn't need it. Furthermore, I think this improves compatibility, because we never know how the format was built (and also think of local patterns). This also open hyphenation to packages (patterns based on rules are easy to program even in TeX, as mkpattern does, for example).
A very complete answer, indeed. You should consider the possibility of publishing it on CTAN (or including it in other manual), including...I wrote an article about it in TUGboat seven years ago, but I've got used to the fact that nobody reads it :-P http://tug.org/TUGboat/tb29-3/tb93miklavec.pdf
Well, but it's not CTAN :-). To find it, people must be aware of its existence. I'll have a look at it. Cheers Javier
