Re: real vs. synthetic width glyphs

Koji Ishii wrote:

>> Actually, I think I do see consensus given the responses on the
>> list. User agents should be *required* to use width-specific
>> variants when the font has them.  This was *not* the resolution on
>> the last call, the resolution on the last call said this should be
>> suggested but *optional*.
>
> Sylvain said he's perfectly fine to allow additional tweaks if doing
> so produces better results for cases such as #12 and fantasia's
> example. Florian is open to do additional tweaks too. It looks to me
> that it exactly matches to the resolution.
> 
> It's not clear to me if you're fine with such a half-open-ended
> approach. If you're, we're all in consensus.

I've stated several times on this list that I'm fine with allowing
wiggle room for user agents in cases where width variants don't exist
for all characters in a tatechuyoko run. But I very strongly feel that
user agents should be *required* to use width variant glyphs if they
are available for all characters in the tatechuyoko run.  This will
lead to consistent results across user agents in the 99% use case,
namely combinations of digits. This was *not* the resolution last
week, the resolution last week left this behavior *optional*.

John Daggett

Received on Tuesday, 9 July 2013 13:25:43 UTC