Ken Whistler wrote: ... > which I think is as faulty as that of people who might claim that, > for example, storing ä for Swedish as <a, combining diaeresis> > would be incorrect from a user's point of view.
I have no problem at all with ä (precomposed) being equivalent to <a, combining diaeresis>. I do have problems with the idea that <a, combining diaeresis, CGJ, combining dot below> should be any different than <a, combining dot below, CGJ, combining diaeresis> (which it formally is). CGJ was unfortunately given the rather special combining class of 0, maybe as a result of the (failed!) experiments at using CGJ as ligator of sorts for combining marks. Invisible as it (now) is (not even affecting combining mark ligation), it does not typographically interact with any other combining mark, and should thus have had a non-zero unique combining class. (I know, it can't be changed now.) I still think CGJ is a kludge through and through, and would have been better off never encoded. Getting back to the suggestion by Ken to actually use CGJ: I would feel much less uneasy about the use of CGJ just before some other combining mark, if that was constrained to certain cases (like, e.g. the Hebrew vowels, perhaps including being preceded by a Hebrew vowel), just like the VSes are constrained to particular applications. (The VSes should also have been given some non-zero unique combining class, but alas. But all vowel marks, not just the Indic ones, should have been given combining class 0, but again alas.) /kent k PS I'm not sure why Ken swiftly classified the suggestion of "HEBREW SECOND VOWEL ..." characters as a kludge... I actually don't see any problems at all with it (the ones listed by Ted Hopp appear very minor, or even non-problems).