font size: proposed plan

To www-style and www-font people:

I have checked the ascenders of letters like b and the descenders of
letters like g in both Times New Roman and Courier New on Windows. It
turns out that the combined height (ascender + descender) is less than
the em height in both cases.

Since many of the browser versions use Times and Courier as their
defaults, I'd rather not change the definition of font-size in CSS to
something that is quite different from the current implementations. I.e.
changing the definition to (ascender + descender) would result in larger
text.

So, I'm inclined to leave the implementation as is for now. Windows
Mozilla uses TrueType's em to select font-size (via negative lfHeight in
LOGFONT). This is also consistent with CSS's current definition of the
"em" unit.

Later, if we find that people want to use fonts that are designed with
very different (ascender + descender) : em ratios, we may want to
consider applying some kind of adjustment (both in the implementations
and the CSS spec).

For example, if a font has an (ascender + descender) height that is much
smaller than the em, then we may want to ask for a larger font size, to
compensate.

My only slight concern with this plan is that if we postpone this
standardization until some later date, it may be difficult to come up
with a clean, simple-to-understand spec that also takes backward
compatibility into account. I.e. it may be early enough now to make an
incompatible change to the spec, but later it would probably be
difficult to make such a change, thereby resulting in a more complicated
spec.

I'd appreciate any comments you have on this.

Thanks,

Erik

Received on Thursday, 27 January 2000 15:40:02 UTC