From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Tue Jan 22 2008 - 10:16:53 CST
Richard Ishida wrote:
> What I'm talking about is what I called transliteration, and defined as a
> method of converting text that allows you to recreate the original source
> from the target (ie. reversability). If you want to do that for a source
> script that is multicameral, you would need some way of capturing whether
> the source contained upper or lower case characters.*
>
> This discussion is exactly why I wrote earlier that I think the
> Transliteration Guidelines document should be more careful in separating,
> describing and labeling these two different approaches.
The guidelines already contains these examples where letter case is aptured
in the transliteration, including for special cases like digraphs, which are
also discussed.
The guideline really says that full reversibility is not possible with
digraphs, because an arbitrary choice has to be made for every
transliteration that contains one-to-several, or several-to-one, or
several-to-several mappings. And the document speaks also about the
existence of different approaches depending on the standard referenced in
the implemented scheme, and says that the "full" reversible schemes are most
often unreadable for native readers of the target script (take the case of
the transliteration of Japanese Kanas to Latin, or the various schemes for
transliterating Han/Hanja/Kanji characters)
This archive was generated by hypermail 2.1.5 : Tue Jan 22 2008 - 11:46:16 CST