2011/9/10 Erkki I Kolehmainen <[email protected]>: > Mr. Delex, > > You should be aware of the fact that the script codes are governed by the ISO > standard 15924 that is produced by the ISO Technical Committee 46. In the > registry for that standard the alpha code Beng and the numeric code 375 > constitute the script code for Bengali. It is only natural (I'd say > mandatory) that the Unicode CLDR documentation refers to international > standards where they exist rather than invent something new and contradictory.
In fact the name of the script used in ISO/IEC 10646 and Unicode predates for long the creation of ISO 15924. The latter just kept the name already used in ISO/IEC 10646. And at that time, there was still no opposition by the Indian or Bengali national bodies about this encoding in ISO 15924 (it's true that there was no requirement for ISO 15924 to keep *only* the name used in ISO/IEC 10646.) In fact, in ISO 15924, it's perfectly possible to have aliases added. As well in the Unicode standard, we could still have property value alias added for the script, not because of its use in specific languages, but due to common confusion or perception where one would think that two alphabets are not unified in the same abstract script. But the problem would immediately explode: how many aliases would we need to register then for the Latin or Arabic scripts? Would these added names, only used for technical purpose, offer any benefit for users? No. You can still speak about the Assamese language and its alphabet, separately of the fact that this alphabet and the Bengali alphabet have been technically unified in the UCS (for good reasons because most of their letters are common by their "genealogy"). The name chosen for the script (between multiple condidates) does not matter because it will still be arbitrary. There must be one which just *unambiguously* designates the correct unified script, independantly of its use in specific alphabets (or abugidas, abjads...). -- Philippe.
