[Spellyans] reDistribution of <i> and <y>

[Michael Everson]

At 00:09 +0100 2008-07-25, A. J. Trim wrote:
>I agree with Owen.

Evidently I don't.

>The <mis>,<res>,<bys>,<bÿs>/<bës> solution has long been the best solution;
>Dan's suggestion (as modified by Owen) has changed that.
>Now the <mis>,<res>,<bìs>,<bys>/<bës> solution appears to be better, and
>needs to be investigated properly.

"Appears to be better?" I say this as an expert 
in orthographic design (a card which I *will* 
play from time to time): it is not very sensible 
to have a system in which a class of words 
*alternate* (as <bÿs>~<bës> do) and yet to only 
mark one of that pair in alternation. How are you 
going to teach this? It is easy to say, hey, when 
you see a diaeresis (which is only used for this 
class of words) you will always know how to read 
the word depending on your preferred dialect. It 
is quite another thing to mark only *one* member 
of the pair. That isn't good othography design.

On top of this, I'm really not prepared to ask 
UC/UCR people to write <gwinn>~<gwidn> for 
'white'. Explaining the <gwin gwynn> distinction 
was easy; they accepted it, and it solved the 
problem of UC <gwyn gwyn>. That happens also to 
be KK practice, and nothing wrong with that in 
this case. But shifting away from the SWF to 
<gwin gwinn> isn't going to get UC buy-in, nor KK 
buy-in, so it's a non-starter.

The people who don't like diacritics aren't going 
to like ANY of them. You can't trade ì/ë for ÿ/ë 
and think that this will make them happy with 
diacritics. The net "reduction" in diacritics 
posited by this proposal is very likely 
negligible too.
-- 
Michael Everson * http://www.evertype.com