Re: How about a typology for input methods

John Cowan wrote:

>
> Peter T. Daniels scripsit:
>
> > > So Arabic, Hebrew, Syriac are not abjads?
> >
> > Unvocalized, they are. Add the points, and they're alphabets.
>
> Most modern Arabic and Hebrew texts, however, are neither fully pointed
> nor fully unpointed: rather, they are strategically pointed with matres
> lectionis.
>
> > > He said he was misled by the way *Unicode* encodes Ethiopic: one code per
> > > sillable, as it does with genuine syllabaries such as katakana or Cherokee.
> >
> > Well, Unicode has nothing to do with script typology! It just does
> > things as they're convenient. Doesn't it have to handle syllabically the
> > Indic scripts that have idiosyncratic treatments of -u, for instance (I
> > think Malayalam is one of them)?
>
> No. All nine Indic scripts are encoded in exactly the same way, and all
> presentation issues are left up to the script engine or font engine.

That doesn't seem terribly helpful.

Lloyd Anderson found, early on, that the nine aren't isomorphic (and you
know how he loves to combine things) -- he even had to devise different
keyboards for both Kannada and Telugu. (Whereas his "Syllabics" keyboard
that works for Linear B and Cherokee and Cuneiform is conceptually
brilliant.)
--
Peter T. Daniels grammatim@...