RE: Complex Combining

[Arcane Jill]

Of course, one really important point is that Unicode text should remain stateless. It would be foolish indeed if, starting from an arbitrary point in the string, one had to parse backwards and forwards to see if there were any invisible brackets. In the extreme, one would have to scan the entire document, and this would be highly undesirable (to say the least). In other words, whilst I do think that invisible brackets are a very good idea, their use should be limited to disambiguating potential ambiguities, and for creating ligated or multi-glyph "base characters" to which modifiers could be applied. Simply put, if a renderer starts from a random point in the string, it should either completely ignore the problem (and ignore any unmatched invisible-close-bracker), or parse backwards by a fixed limited number of characters. This would effectively prevent the use of invisible brackets as "fake markup".

Presumably, the same problem (that of remaining stateless) is created by the music characters (e.g. U+1D173..1D174 MUSICAL SYMBOL BEGIN BEAM..END BEAM). That is, if a renderer starts in the middle of the string and encounters an END BEAM, how far back is it required to parse in order to find the BEGIN BEAM before deciding that it's unmatched?

Just my thoughts. Despite the objection I just raised, I still like the invisible brackets idea, and still think it's a good one.

I further think that a renderer ought to be allowed to interpret <invisible open bracket><some musical notes><invisible close bracket><musical symbol end beam> as completely equivalent to <musical symbol begin beam><some musical notes><musical symbol end beam>. This would be a start toward a consistent approach, whilst also retaining the legacy method.

Jill


 -----Original Message-----
From:     Philippe Verdy [mailto:[EMAIL PROTECTED]]
Sent:    Saturday, November 29, 2003 2:21 AM
To:    Peter Kirk
Cc:    [EMAIL PROTECTED]
Subject:    RE: Complex Combining


I think it would be simple to have invisible parentheses in that case, and be able to apply the diacritic in the group:
<invisible open bracket><diacritics><one or more characters><invisible close bracket>.