RE: Transliteration

I think we can expand on more complex schemes for denoting
transliterations as much as we want. But it gets more and
more complex. It may be that a level of indirection is needed:

- Some tag, maybe an extension of language tags, is used for
	identification. E.g. ja-hepburn, ja-kunreisiki, and
	so on.

- The registry of such tags, maybe the language tag registry,
	could provide additional information, if possible in
	machine-readable format.

I think that this might often be a better approach, because it
is shorter and it allows for well-known names. For examlpe
ja-hepburn would have to be clasified as

	source-langugage:		Japanese
	target-language-consonants:	English
	target-language-vowels:		Italian/German

For ja-kunreisiki, it may look like:

	source-langugage:		Japanese
	target-language-consonants:	Artificial/Japanese
	target-language-vowels:		Italian/German

Regards,	Martin.