I may give some excuses to him, if he is not aware of the technical justification of why names are immutables. But what he really wants is to avoid being exposed to these "Bengali" names. This is not a matter of tehcnical encoding, but more a question of localisation (for example when using a character picker application, or when searching character collections by names).
Nothing forbids a localized application of using accurate names that match a specific language expectation about its alphabet. But Mr Delex must understand that the UCS (by Unicode or ISO/IEC 10646) does NOT encode language-specific alphabets, but "unified" scripts that share a lot of common letters and a common structure is such a way that those languages can be freeely mixed and interchanged without duplicating the letters. May be the question could be forwarded to the CLDR project, about the localisation of letter names for language-specific alphabets. For now the CLDR project still has problems in just knowing which letters are representative of the orthography used a single language (for example, is the letter "é" is part of the English alphabet ? Is the letter "ā" part of the French alphabet, because it is used in official toponymy ? Same thing about "Å" for example in "Åland"...) Just consider how we use the alphabets today, we frequently borrow foreign letters from foreign alphabet, very easily because they are in fact part of the same unified "script". Still, we do not need to necessarily locally name those borrowed letters using the name of our local alphabet for out local language. But new characters won't generally be reencoded in the UCS (the UCS still chose to NOT unify the Latin, Greek and Cyrilic alphabets, and even chose later to desunify the Coptic alphabet from Greek; on the opposite, it refused to desunify the Fraktur and Celtic alphabets from Latin, because there was no frequent cases, clearly contrasting, where such desunification would be necessary; at the same time the UCS still maintains the IPA symbol set as a full part of the Latin alphabet, even if it required reencoding in the Latin script some Greek letters). There are tehnical tradeoffs in those decisions of unification or desunification of scripts. But it is important to understand that scripts in the UCS are definitely not the same as alphabets; scripts still need to be (arbitrarily) named from the name of the most representative alphabet encoded in it (or an alphabet already supported by a widely used legacy standard), and there's a good reason to give technical names for all characters encoded in that script, that reference this arbitrary script name, independantly of their use in language specific alphabets. Notes: - ignore above the subclassification of "alphabets" into "true alphabets", "abjads", "abugidas", or even "syllabaries", even if this classification plays a very important role in the decision of unifying them or desunifying them in the same "script" in the UCS. - For Mr Delex: the "UCS" is the Universal Character Set, i.e. the same **unified** repertoire standardized and encoded internationally by both the Unicode standard and the ISO/IEC 10646 standard (and all their annexes). Both standards do not encode language-specific alphabets and cannot even give distinctive names used in various languages to reference the same unified characters. -- Philippe. 2011/9/14 Erkki I Kolehmainen <[email protected]>: > Dear Mr. Delex, > > Please, please spare us from further details in support of your crusade. You > should finally accept the fact that the official block name cannot be > changed, the rest is effectively OT. > > Thank you! > > Erkki I. Kolehmainen
