Re: superiors, inferiors, ordinals, etc. (was: [CSSWG] Minutes and Resolutions 2010-03-17)

Christoph Päper wrote:

> If ‘vertical-align’ is either ‘sub’ or ‘super’ we can guess the proper
> transformation to do, so authors should only have to specify ‘auto’.
> 
>   character-transform: _normal_ | auto   (inherited)
> 
>   ‘normal’:  -subs -sinf -supr -ordn / ‘No Vertical Position’
>   ‘auto’:    +subs +sinf / ‘Inferiors’ where “vertical-align: sub”,
>              +supr +ordn / ‘Superiors’ where “vertical-align: super”

I don't think it's such a good idea to key the behavior of one
property on the value of another, it really hides the meaning of the
property.  I realize the definition I proposed also has side effects
on other properties but it's minimal and only for compatibility sake.

> PS: I’m probably ignorant or lazy again, but why is this not a
> ‘font-variant-*’ property?

No, that's a perfectly reasonable question.  The reason is that
subscripts/superscripts are semantic in nature, a
subscript/superscript style should result in subscripted/superscripted
text, whether or not the font supports subscript/superscript glyphs as
a feature. For other font-variant-* properties the fallback is always
just the default glyph when a feature is not supported, the idea being
these reflect a stylization and not something that alters the content.
Lots of gray areas here but I think subscripts/superscripts clearly
are part of the content and need to be handled differently.