Re: Baselines (was RE: Units, font sizing, and zoom suggestion for CSS 3)

Bert Bos wrote:
> 
> There's currently text in the works in the CSS WG to specify these
> various baselines.

Thank you very much for sharing this on the open mailing list! I
sincerely hope we will continue to see such sharing of info.

> The definition of 'vertical-align' may have to be
> re-interpreted to deal with texts that mix scripts.

Yes, I think so too.

> For a formula and a Japanese quote, you could set, resp.:
> 
>     MATH { baseline-identifier: mathematical }
>     Q[LANG|=jp] { baseline-identifier: ideographic }

It's nice that you give an example using the Q element and the LANG
attribute, but I think we need to keep in mind that real world documents
often mix ASCII and Japanese characters, without the use of Q and LANG.
We need a baseline story for this too.

> (Another name for 'lower' could be 'latin', because 'ideographic' is
> actually lower than 'lower'...

Hmmm, that's interesting. I don't know whether ideographic would be
considered lower than "lower". Ken Lunde's book CJKV Information
Processing says that East Asian characters are "optically centered"
rather than resting on a baseline. If you look at MS Gothic on Windows,
for example, you will see that the baseline cuts through the bottoms of
the glyphs.

> In combination with this, there is also a proposal for a
> 'line-height-policy' property that can, among other things, enforce a
> mode in which the line height throughout a paragraph remains constant,
> despite text with different baselines or sizes.

That sounds similar to the min-line-height and line-height-override
stuff that we've been discussing at mozilla-layout@mozilla.org.

Erik