Re: [CSS21] 3 editorial modifications in § 15.5 Small-caps: the 'font-variant' property

On Wed, Oct 12, 2011 at 05:39:24PM +0200, Chris Lilley wrote:
> On Tuesday, October 11, 2011, 2:42:46 AM, Gérard wrote:
> 
> GT> # It is acceptable (but not required) in CSS 2.1 if the small-caps font is a
> GT> # created by taking a normal font and replacing the lower case letters by
> GT> # scaled uppercase characters.
> GT> 
> GT>   -- section 15.5 Small-caps: the 'font-variant' property
> GT>      http://www.w3.org/TR/CSS21/fonts.html#small-caps

> Both letters and characters are incorrect here. Firstly, one can't scale a
> character but one can scale a glyph. Secondly, it could be misread as an
> actual substitution of characters (which would show up in the DOM).
> 
> | It is acceptable (but not required) in CSS 2.1 if the small-caps font is
> | created by taking a normal font and replacing the lowercase glyphs by
> | scaled uppercase glyphs.

(Preface: I'm ignorant both of small caps usage in different languages
 and of usual OpenType handling of things like small caps and the
 diacritics involved in the transformations referenced below.)

I'm not sure that this is quite right either: I'd have thought that we'd
still want the transformations in Unicode's SpecialCasing.txt file[1] to
apply (and note that some of these transformations are context-dependent)
rather than trying to apply a purely single-glyph-to-single-glyph mapping
that the above proposed wording suggests to me.

[1]: http://www.unicode.org/Public/UNIDATA/SpecialCasing.txt

So I'd have thought that we still want a string-to-string transformation
to be involved, so that context-sensitive transformations happen
correctly, though of course that's not to say that 'font-variant' should
affect the text visible in the DOM.

Perhaps the main point that the spec should convey is just that it's
acceptable to use scaled uppercase glyphs in place of true small-caps
glyphs: it needn't say anything about how to choose the string of glyphs
(as any mention of lowercase glyphs tends to imply) if we don't want to
specify the handling of cases such as given in the SpecialCases.txt file.

pjrm.

Received on Monday, 17 October 2011 01:26:43 UTC