Just out of curiosity, I'm wondering what those here think about unicode and, in particular: 1. Is its concept of character -- basically unsigned 32 bit integer -- durable for, say, the next 100 years? (As I read the discussion here, I think not.) 2. Do we think that 2^32 is a wise upper bound? (This question vanishes if we think that representing characters as integers, rather than as more complicated data structures, is inadequate.) Unicode is directly relevant to the future of LaTeX to the extent that LaTeX is going to be robust for formatting XML document types because normal document content can consist of arbitary sequences of unicode characters. XML systems are designed to make decisions only where markup occurs. It is reasonable for an XML processor writing in a typesetting language to know the markup ancestry of a character, e.g., whether it is within a math zone, but not reasonable -- unless the processor, like David Carlisle's xmltex, is a TeX thing -- for it to know that a particular character must have \ensuremath applied. I note that in GNU Emacs these days characters can have property lists. Thanks for your thoughts. -- Bill