--- In qalam@yahoogroups.com, "suzmccarth" <suzmccarth@...> wrote:

> I also wonder what the Vai SIL keyboard uses,
> surely some other roman orthography.

Transliteration on the fly. Differences from the Unicode character
names are:

1) Lower case.
2) /e/ is 'e', /E/ is 'q', /o/ is 'o', /O/ is 'r'.
3) Egressive /d/ is 'd', implosive /d_</ is 'dl'
4) Vowel nasalisation is a preceding tilde (~), not a trailing 'N'.
5) /N/ is 'xg~', not 'NG'. Curious! NGGEN is described as 'other',
without any hint as to its pronunciation.
6) /S/, /Z/, /T/, /D/ are not supported, so no SHA for the book of Ndole.
8) The keyboard does not suppport /c/; I'd expect the implied
transliteration to be 'c'.

The keyboard only works for the characters they could squeeze into the
8-bit SIL Vai font - the overflow characters in SIL Vai Extras have to
be entered by other means. I was wondering how Michael Everson would
handle a font-sensitive encoding. (I know it's not impossible to
handle it.)

There are some odd-seeming encodings:

HIN is '~i' (as in Tucker)
HAN is '~a'
NGAN is 'h~u'
AN is 'xg~a'
ON is 'xg~r' - I've already mentioned the vagaries of this symbol.

NGON is described as /hO~/ (implying 'h~r').

Richard.