ISO 10646 is designed to be what I will call "full
multilingualism". That is it supports "all" language or scripts [...]
That was certainly the expressed original design goal.
(with
the exceptions of some rare or little used ones which have not be
standardized enough to include [...]
And this, plus the issues raised by unification of some things that
arguably should not be unified, plus the issues raised by the handling
of composed characters without uniqueness-generating composition rules,
are the sources of the arguments that, relative to that original goal,
10646 is an unsatisfactory realization.
[...] yet.)
So, what are we waiting for? And when is it likely to happen? If
10646.n, N>1 is needed to have a true UCS, then these discussions are
premature at best, since there are no real proposals on the table for
things off the BMP. And we are in UCS-4 -land.
Once again...
-- There things that 10646 demonstrably does well. This is not a null
set.
-- There are things that 10646 demonstrably does poorly. This is not a
null set either.
-- There are ways around many of the things that 10646 does poorly.
Some of them involve profiling and/or explicit language specification.
Others require modifying or extending 10646.
Since the second category is non-null, we must ask whether:
(i) 10646 provides enough value-added in the first category to justify
a transition from other strategies. And either...
(ii) We are willing to not solve the problems in the second category
or...
(iii) We are willing to put in the necessary pain, effort, and
suffering to do the work needed to make it useable for our purposes.
It is not clear that these religious wars help very much with any of
this.
--john