From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Fri Apr 14 2006 - 19:14:54 CST
From: "Peter Constable" <petercon@microsoft.com>
> Your entire argument here seems to me to be built on the assumption that the average
> user wants to represent their characters in multiple ways: as a single character here
> and as a character sequence there. I've yet to meet a user like that. This argument
> simply isn't valid. What *would* have been a valid argument supporting a decomposition
> mapping would have been evidence that the text element in question is sometimes
> written as two discontiguous graphic elements resembling the combination of 1B0D and
> 1B35, but apparently that does not happen.
What will happen in Unicode if it actually happens in Balinese texts, where both drawings would be considered equivalent? It seems that the ligature is very opportunistic for the handwriting, due to the way that 1B0D is written (so the drawing pen just continues on the samestroke to terminate the tedung). But I wouldnot bet that this never happens in artistic variants.
This means that 1B0D 1B35 may happen, and it will be considered distinct (in Unicode only,but not for readers) from the encoded ligature. I just want to get sure that there's a strong advice by the Balinese community and experts that this difference will be significant, and that 1B0D 1B35 would be considered not only as a bad typography, but as a orthographic error (including for artistic renderings, with ornaments).
Anyway, what would be the UCA collation sequence for both sequences? Isn't this proximity already part of the normal ordering for Balinese (at primary level if encoded separately, or secondary level of this proximity is reflected in the normal sort)? This is what would be much more conclusive and convincing than just the relative occurences of glyphs discovered.
This archive was generated by hypermail 2.1.5 : Fri Apr 14 2006 - 19:16:35 CST