From: "D. Starner" <[EMAIL PROTECTED]> > Unicode will not allocate any more codes for characters that can be made > precomposed, as it would disrupt normalization.
But what about characters that may theorically be composed with combining sequences, but almost always fail to be represented successfully? If normalization stability is a legitimate goal, this should not prevent some precomposed characters to be encoded separately without any canonical equivalence with their logically decomposed equivalents. This is already true for combining overlay diacritics (single or double strokes, vertical bars, and horizontal bar), or for diacritics that form a ligature with the base letter (such as combining horns and cedillas...), where the logically "composed" characters should better have their own codepoint. If such ligature has a distinct semantic from a ligature created by ligaturing separate letters for presentation purpose, the character is not a ligature (the AE and OE "ligated glyphs" are distinct abstract characters) . The case of dot below however should be handled in fonts by proper glyph positioning and probably not by new assigned codepoints, unless this is only one possible presentation form for an actual distinct abstract character that may have other forms without this separate diacritic (for example if g with dot below was only one presentation for an abstract character that may be also renderd with a small gamma).... We have already some examples of characters that have exactly the same glyphs but are distinct semantically due to their case mappings (see the glyph showing a barred D for example, or the turned E: their apparence may be identical but they have distinct semantics and properties).
