Re: Why doesn't 'font-weight: 100' work yet?

Bert Bos:

        CSS 3    Fontconfig    Gill S.   CSS 3 example names / algorithm
  100            thin          Light     ^
  200            extra light   Light     ^
  300            light         Light     ^
  400   normal   book          Regular   Book, Regular, Roman,  
Normal, Medium
  500            normal        Regular   ^ Medium
  600            demi bold     Bold      v
  700   bold     bold          Bold      v Bold
  800            extra bold    Bold      v
  900            black         Bold      v


Btw., the algorithm in the current (quite old) draft of CSS 3: Fonts  
requires to try to find something darker than '900', when it's not  
available directly. That always returns false as far as I understand.  
(There's also a "th[a|e]n" typo in the beginning of that paragraph.)

> (Fontconfig uses slant instead of font-style and  
> instead of
> 100...900. We could discuss whether "book" maps to 400 or to 500, but
> for the rest the mapping is straightforward.)

Do you mean we should discuss whether 'book' should come before or  
after 'normal'?

I don't know anything about fc-match and hardly more about font  
weights in general, so the table above, which I assembled from your  
data and the WD, appears strange to naive me:
- Why do nine steps map 3:2:4 to three available "styles", instead of  
- Why isn't "book" called "demi light" (or "demi bold" something else)?
- Why is book-style apparently not quite normal?
- Why doesn't CSS 'normal' match Fontconfig's?
- Why are there only two absolute keywords in CSS?
   ("Black" is already used for colours, but that shouldn't matter,
   neven in the 'font' shorthand property.)
- Why are there unit-less numbers instead?
   (Those are sometimes frowned upon elsewhere.)
- Why are they '100'-'900', not '1'-'9' or '0.1'-'0.9'?
- Why are they not expressed as percentages?
- Why is there no '0' (no ink) and '1000' (all ink)?
   (Pretty useless, but closer to established numeric values elsewhere.)
- ...

Received on Wednesday, 22 August 2007 23:07:06 UTC