From: "Michael Everson" <[EMAIL PROTECTED]> > At 20:43 +0200 2004-08-21, Philippe Verdy wrote: > > >So for now, Peter has listed these strategies, but ignored the one > >consisting in encoding only a new undecomposable HOLAM MALE base character > >(I know that this was the first idea that SII rejected several years ago, so > >I doubt that it would accept it today, unless SII becomes convinced that it > >is the only viable solution to limit the various interoperability and > >security problems that any disunification of the existing HOLAM DOT would > >create). > > This would be encoding a DUPLICATE character and will not sneak > through WG2. I promise.
Why would that be a "duplicate"? The existing HOLAM MALE compatibility character is not really a HOLAM MALE, as it has the wrong canonical decomposition, meaning that it just represents the pair of characters VAV and HOLAM, and it is also excluded from recomposition (so the existing HOLAM MALE compatibility character will never exist in texts in any normalized form.) And its interest is clear: it avoids the multiple encodings that may persist when a new HOLAM DOT will be introduced, as this new HOLAM may appear after any base character (not only VAV), due to implementations of editors or keyboards extended to support a new HOLAM DOT. If least cost is intended, then the other alternative is to encode VAV HALUMA separately but it is rather illogical as it is clearly two abstract entities, including phonetically. This logical consideration is very unlike HOLAM MALE, which is just a single vowel where the base character is just there as a needed silent holder to write an isolated HOLAM vowel, or to be used in unpointed texts to mark the presence of an additional syllable (where the presence of a vowel is implied, and interpreted as HOLAM only by the reader as the vowel is not encoded or rendered): VAV is used as this silent holder because it is not a strong consonnant (in HOLAM MALE, it plays more or less the role of the half-consonnant W in the modern Latin script, where W is also sometimes read as a O or U vowel; or it plays a role similar to the Roman Latin letter V which is both a vowel U or consonnant V, and also silent in some cases, or acting as a phonetic modifier for a preceding vowel A like in AVRORA).
