Ken, thanks for your empathy! Actually, our programmer is experimenting as
we speak with Norm Walsh's DTD flattener program. I'm testing the output
right now. If it works, we can do the groovy modular DTDs and just flatten
them before storing them in the repository. Of course, I don't think we can
then unflatten them properly, so we'll still be stuck having two versions of
our DTDs - a flat one for use by documents in the repository, and an unflat
one as a master. Oh well, it'll work for now.