> So, in effect the UNICODE character names attempt to be > a unified transliteration scheme for all languages? Are these > principles laid down somewhere or is this more informal?
The Unicode character names attempt to be (a) unique and (b) reasonably mnemonic. Anything beyond that is a bonus. They expressly do *not* represent any form of transliteration or transcription scheme.
However, it is sometimes forgotten that the standard is intended to be in English (with the possibility of translation to other languages, for example the French translation that has been carried out for 3.2).
If a character has an obvious or common English name, that name should be used. Where there is no obvious English name, a transliteration or transcription of the native name makes sense.
In the case of a script used by multiple languages, it's an interesting question which language wins out. Assume you have a majority language that doesn't use a certain character, but has a word for it. Does it make more sense to keep all transscriptions in the same language in Unicode character names?
Opinions differ.
Ultimately the only strong requirements are that names are unique and (recently added) that dropping common words such as LETTER, MARK, SIGN and SYMBOL as well as spaces, and hyphens do not affect that uniqueness.
Since the character names freeze mistakes permanently and since the committee decisions have resulted in some odd and not always consistent approaches to naming, some of the translated sets of character names are more consistent and usable than the official English.
That has led to the suggestion of eventually creating a translation of the character names into e.g. American English, essentially providing a set of consistent aliases that might be useful for dictionaries of character names exposed to end users interested in locating characers, as opposed to merely wanting the formal, but potentially arbitrary reference.
A./