Re: [CSSWG] Minutes F2F 2009-06-05 Part I: SVG Properties, Fonts

Thomas Phinney <tphinney@cal.berkeley.edu> wrote:

> Just one comment:
> 
> ...
> >  Steve: Original spec implied that values like 250 could not be
> > mapped into CSS model.
> >  John: It is a 9-point scale, forget about the actual values. Just
> > need to map the font onto nine points, whatever their names.
> 
> No, it is NOT a 9-point scale. It is a numeric system of ~1000 units.
> Treating it as a 9-point scale (or as arbitrary labels) will get you
> into trouble.

The font-weight property *as currently specified in CSS* is a 9-point
scale with arbitrary, ordered labels, that happen to have the form of
numbers. Quoting http://www.w3.org/TR/CSS2/fonts.html#font-boldness --

'font-weight'
  value: normal | bold | bolder | lighter | 100 | 200 | 300 | 400 
  | 500 | 600 | 700 | 800 | 900 | inherit

If it were a numeric system it would be

  value: normal | bold | bolder | lighter | inherit | <integer>

Now if what you are saying is that in order to handle all currently-
existing fonts, the spec needs to be changed to allow any integer in
the 0 ... 999 range, then OK, make that proposal.  But that is not the
way it is now, and I for one would want to see an example of a font
with more than nine weights in order to be convinced that a simple
mapping from this 9-point scale to whatever the values are in the
actual font would not suffice.  I'm skeptical about this because the
text at http://www.microsoft.com/typography/otspec/os2.htm#wtc leads me
to believe that there are exactly nine possible values for
usWeightClass in a valid OpenType font.  (Would you be happier if the
spec were changed to use the names 'thin', 'extra-light', ... 'black'
from that list, instead?)

zw

Received on Wednesday, 17 June 2009 19:56:13 UTC