--- In qalam@yahoogroups.com, "Richard Wordingham" <richard@...> wrote:

> I presume there was no choice as to how a CVCCV sequence should be
> broken up into orthographic units, because I can't see how CVC.CV and
> CV.CCV would be encoded separately in Unicode. Unicode encodes CVC
> with the 'virama' *after* the final consonant, rather than before it.

It was rather late at night when I came to that conclusion. The
answer, of course, is that a CVC.CV rendering would use the ZWNJ
character. The only issue is whether this is just common sense or is
actually specified by the Unicode standard.