Re: [css-text] Amending hanging-punctuation for Western typography

On Thu, 2016-04-07 at 00:41 +0000, Alan Stearns wrote:
> On 4/6/16, 5:34 PM, "fantasai" <fantasai.lists@inkedblade.net> wrote:

[...]

> >  You just have one of the possible "glyphs" in the pair be SOL or
> > EOL.
> 
> Ah! Yes, that could work - and we’ve entertained the idea of letting
> authors define pair-kern data before. But I think that would have to
> be font-specific, and apply only when the font matches for both items
> in the pair. So that quickly becomes complicated, too :)

I don't really see how StartOfLine or EndOfLine could be in a different
font than the character being "kerned". I've encountered this SOL/EOL
concept before under the name of Margin Kerning.

Explicit kerning in the font metrics should be just an override though,
because existing fonts don't have it. So a default of, e.g. a
percentage, or a list, (character percent character percent...) for
start of line and for end of line, where 0% is the default and 100%
means "sticks out entirely into the margin" and 33.333% means sticks
out a third of its width, would be manageable in practice I expect, and
would allow different defaults in different languages.

I don't think there's much value in sets of characters for Western
typography. The ones you want to hang are ( ) , . - and minus, en dash,
quotes, and the percentages probably vary. A single default value for
start and another for end would go a long way, and the actual list
would let tools emulate Illustrator in specific locales/languages. 

Also, kern pairs are usually defined on a per-glyph basis, since it's
the glyph that hangs, but I think we don't expose glyph names or font
encoding positions elsewhere. I'm not sure if that would matter in
practice, although it'd mean you couldn't hang the "quaint" ct ligature
c_t into the margin.

-- 
Liam "Evil XML" Quin <liam@w3.org>
The World Wide Web Consortium (W3C)

Received on Thursday, 7 April 2016 03:03:50 UTC