You seem to under-estimate the requirements for compatibility with ISCII and
at the same time over-estimate our ability to break new ground. The need for
compatibility in translating existing data from ISCII to Unicode and back in
as straightforward a manner as possible (i.e., one-to-one linear mapping) is
extremely high.
In any case, as Ken Whistler has noted, Devanagari is one of the few scripts
in the standard that closely follows the ideal principles of the standard (at
bottom, this means don't encode presentation variants and encode only logical
entities). And here, half-consonant forms and conjunct forms are judged to
be presentation variants predictable from an underlying logical representation.
If you can provide overwhelming evidence that these forms do represent graphemic
distinctions (as opposed to merely allographic), then it is possible that your
desires could be met. In the mean time, accept what is there as a given.
Regards,
Glenn Adams
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:33 EDT