Re: [CSS21] last edition: pity

> I think WPF name matching system is a very interesting example of how
> hard it is to map real-life font families onto the Procrustean bed of
> CSS font families.

The faults you and others have pointed out seem to have more to do with
older API's like GDI, rather than a specific fault of the model per se. 
Font weight values in CSS are based on the weight values defined in
OpenType:

  http://www.microsoft.com/typography/otspec/os2.htm#wtc
  
If GDI problems force vendors to ship with skewed values (e.g. 250 ==>
Thin, 275 ==> UltraLight, etc.), that's a complication but not something
that necessitates a change in the model.  Implementors will lose a
little more hair but it's not something that needs to get pushed through
to authors; specifying a thin face should result in a thin face if one
is available.

Do you have suggestions for a better model for weights?  Do you feel CSS
font weights should be any integer value between 100 and 900?  Or that
simply more weight values should be allowed, a 13-point scale instead of
a 9-point scale for example?

John