Re: note about disabling font-variant

Koji Ishii wrote:

>> # Full-width glyph variations can also be chosen via the
>> # font-variant property or @font-face descriptor [CSS3-FONTS].
>> # However since such variations do not affect the glyph shape in
>> # contemporary fonts, there is no explicit interaction with this
>> # feature currently. A future revision of this specification may
>> # require ignoring such declarations within the combined text if
>> # common usage makes such behavior desirable. 
>> I strongly object to this wording.  Who is this note addressed to
>> and what purpose does it serve?  You're conjecturing about possible
>> future WG decisions and as such I don't think that has any place in
>> a spec. It's just noise to those working on implementations and
>> distracting to authors trying to understand the
>> 'text-combine-horizontal' property.
>> Either omit this note or replace it with one directed at authors:
>>   Properties that affect glyph selection, such as the
>>   'font-variant' and 'font-feature-settings' properties defined
>>   in [CSS3-FONTS], can potentially affect the selection of
>>   variants for digits or alphabetic characters included in
>>   <em>tatechuyoko</em> text runs. Authors are advised to use
>>   these properties with care when 'text-combine-horizontal' is
>>   also used.
>> That's more informative and serves a purpose I think.
> How about combining the two? Your text is better directed for
> authors, but it's hard to understand how it "can potentially
> affect." Fantasai's text can explain that.

That note doesn't explain anything about the interaction and it misses
the point that the automatic disabling of full-width variants is
generally unecessary because the design of full-width variants
generally doesn't vary much from the default design *and* other width
variants usually take precedence over full-width variants.

> If I understand correctly, didn't you mention (or agreed?) that if
> fonts that are affected by the removed wording, such as Adobe Ming
> Std, become majority, we may revisit the discussion? It looks to me
> that the text describe such situation very well.

Revisit the discussion, yes, add possible future normative
requirements in the form of a note, no.

The details of Adobe Ming Std again illustrates my point.  Like MS
Mincho, the glyphs for the default digits and half-width variations
are fixed at half-width.  The glyph extents of the full-width digit
glyphs are almost precisely those of the default digit glyphs and the
supported width variants of Adobe Ming Std override the use of
full-width variants when both are specified. So there is no need for
automatic disabling of full-width variants.

I think to justify normative requirements for implementations, such as
"disable full-width variants when text-combine-horizontal is used",
there needs to be a reason.  Unless there are both actual fonts and
real use cases, we should omit such unnecessary requirements from CSS
specs.  In addition, it would probably make sense to focus on the use
cases for *Japanese* fonts, since the primary use of tatechuyoko is
within Japanese publishing and Adobe Ming Std is a Chinese font.

Given the late date, I think the most expedient solution would be to
omit this note and not haggle about the wording of a note that
describes an edge case.

John Daggett

Received on Tuesday, 24 September 2013 01:39:43 UTC