Murray said: > MathML does have markup to extend diacritics across arbitrary numbers of > characters and it's not likely that MathML would use the CGJ for this > purpose.
...which implies that we would also need to make an addition to UTR #20, Unicode in XML and other Markup Languages, before the UTC could really approve the concept. Another reason to go slow and make sure all the angles are deliberated (and trisected) first. > But it would be handy for representing such expressions in > plain-text Unicode. I agree -- as for the dictionary example that started this thread. However, it might make sense to make an implementation guideline that would constrain any such mechanism to double diacritics and suggest that people move to generic markup mechanisms if they need more. Thus: X CGJ X CGJ combining-breve But not: X CGJ X CGJ X CGJ combining-breve etc. This would keep the expected plain text scope down, and would minimize the size of the ligation entries in fonts -- probably a good idea for performance -- without significantly cutting down the usefulness of the mechanism for plain text. My guess is that well over 99.9% of the use of these beasties in what could arguably be called plain text is application of diacritics over/under a digraph. Application over or under a tri- or higher-order multigraph must be very rare indeed -- in part just because anybody trying to use such things in print (or on typewriters, or whatever) would have run into technical difficulties and other pushback in the first place. --Ken
