Re: [css3-fonts] font-variant-numeric

fantasai wrote:

> 1. Grammar is too loose.
> 
> The current spec specifies
>   # <numeric-values> = [lining-nums | oldstyle-nums | proportional-nums | tabular-nums
>   #                    | diagonal-fractions | stacked-fractions | slashed-zero]+
> 
> But some of these pairs are mutually exclusive, and there's no reason any one
> keyword should be specified more than once. The grammar should be
> 
>   | <numeric-values> = [ lining-nums | oldstyle-nums ] || [ proportional-nums | tabular-nums ]
>   |                    || [ diagonal-fractions | stacked-fractions ] || slashed-zero

Yes, I agree this is nicer, as Christoph has pointed out.  However, the
one problem I see with this is that font-variant is still a shorthand
that could take multiple <numeric-values>.  So the usage below would be
handled inconsistently between the 'font-variant-numeric' and the
'font-variant' shorthand:

  font-variant:         lining-nums tabular-nums slashed-zero oldstyle-nums;
  font-variant-numeric: lining-nums tabular-nums slashed-zero oldstyle-nums;

With your definition, the second one would be valid syntax, the first would not.

> This makes the following sentence redundant:
>    # The values ‘lining-nums’ and ‘oldstyle-nums’ are mutually exclusive,
>    # as are ‘proportional-nums’ and ‘tabular-nums’, ‘diagonal-fractions’
>    # and ‘stacked-fractions’.
> It should therefore be removed.

Nope, this is still needed because @font-face font-variant settings
and general property font-variant settings have to be resolved, see
section 7 of the spec.

> 2. Keyword definitions too terse.

Yes, I was waiting until I felt there was agreement on the set of
features before adding more descriptions.

> a) Several of these terms have reasonably-common alternatives, which
> should be added in parentheses so that people looking for this feature
> can find and recognize it easily.

I'm not at all clear what you mean here.

Received on Wednesday, 24 March 2010 05:51:17 UTC