Elaine asked: > Dear John Cowan: > > > Instead, the alefs will sort together, the bets will > > sort together, and so on. > > Only if two words are identical in everything but > > script will the Square > > come first and the P. second (or the other way > > about, depending on how > > the default collation rules are set up). > > So could you do this with all Semitic/Afroasiatic > languages which have something like alef and beth?
Yes. > Is > there a numeric limit? In principle, no. In practice, yes. However, in practice, the principled limit on the practice far exceeds the practical need. ;-) If there are 2 Semitic/Afroasiatic scripts encoded, or eventually 4 (or 9), then all the alefs and bets *could*, in principle, be weighted together, which would impact both searching and sorting. If 97,621 Semitic/Afroasiatic scripts were encoded, then chances are most actual implementations of this would blow up on some numerical limit. > Or if the Egyptian > biconsonantal etc. stuff is harder to process, is that > a limitation? No. You just weight the biconsonants as units or the units as having two weights, depending on what effect you are trying to get for the search and/or sort. --Ken > > Elaine
