From: Patrick Andries (Patrick.Andries@xcential.com)
Date: Sun May 16 2004 - 11:07:56 CDT
Peter Kirk a écrit :
> Well, at least "façade" and "facade" collate together at the top
> level, with the default collation weights, and so one will match the
> other in simple searches.
[PA] I was simply trying to say -- not that I always express myself well
-- that adding some characters may force additional processing (here in
the collation, elsewhere if a cedilla exists as a combining character
in normalisations and rendering). Adding characters is not as innocent a
process as some seem to say : «We just add characters and that's it, you
are not forced to do anything about it». If it is true that one is not
forced to use them as a writer in the script, when one does not control
the writers or sources and one has to process several sources (collate,
render, search them), one is then forced to implement certain additional
processes (for excellent reasons if the characters are indeed
necessary). This is why I believe one must carefully review the pros and
cons before adding new characters, they may well be unified with
existing ones, for example.
> Again, if the separate Punic script were to be compatibility
> equivalent to Phoenician or Hebrew I would not have strong objections;
> but otherwise I am sure that there would be strong objections on the
> grounds that yet further splitting of what is logically the same
> script used for closely related languages leads to even more confusion.
[PA] I would have like Michael to say that splitting may lead to
confusion with little gain..since he suggested ths unification.
Note that I believe unification of Neo-punic with Phoenician is the
prudent course to take (for the reasons I explained : introducing new
characters has a cost and does force people to do something about them).
Otherwise, if Unicode has space, tailoring collations is The Proper
Thing To Do and «Unicode doesn't force people to do anything. Unicode
makes characters available for those who wish to use them. », why not
encode Neo-Punic ? After all, one could make a case for it : Neo-punic
is a remote descendant from "Canaanite" (genealogically as much as the
Aramaic-Square Hebrew branch, it also retains the 22 primitive Canaanite
characters), pretty different as far as glyphs are concerned (some
simple strokes may represent a "b", a "d" or an "r", a Saint-Andrew's
cross may represent "m" or alef), has three subcategories (Carthago,
"Tripolitaine" and "Maghrebine"), some inscriptions (cf. Cherchell) are
mixed Neo-punic and Punic (how would one represent them in plain text?),
it uses matres lectionis (reusing gutturals having nearly completely
disappeared in the spoken language), etc.
This archive was generated by hypermail 2.1.5 : Sun May 16 2004 - 11:08:42 CDT