From: "Peter Jacobi" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, November 10, 2003 9:51 AM Subject: Re: Tamil 0BB3 and 0BD7
> Hi Doug, All, > > "Doug Ewell" <[EMAIL PROTECTED]> wrote: > > [..] Second, disunifying "y" > > would cause untold mapping nightmares. [..] > > Not exactly nightmares, but the Tamil case does > cause some mapping discomfort. > > About three out of five 7bit/8bit encodings for Tamil > have these two Unicode codepoints unified. And at > least one of the dis-unified ones did the dis-unification > exactly for Unicode interoperability. So it remains at least one of these encodings where the characters were disunified. This should have been solved in Unicode by creating a compatibility equivalent (so that bijective mapping with this encoding remains possible), but with a compatibility decomposition to the other code point (like for CJK compatibility ideographs that had two position assignments in legacy encodings). > So in effect, Unicode handling of this case, may actually > change Tamil use - I've already seen proposals to > a script reform dis-unifying the glyphs. I bet that the AU length mark should have been this compatibility assignment in Unicode, because the same letter can be treated individually as a base character, and should prevail to the combining interpretation which is language-specific and contextual, butnot needed for the correct rendering of Tamil. But yes, this separation in code points will have the effect of users wanting to create glyph variant for the AU length mark, finally splitting the usage of the letter the same way as it occured in Latin between U and V, simply because it allowed easier reading by removing the required contextual analysis of the written glyph.
