CSS Draft 8/95: prioritised fount lists
Thore Husfeldt writes:
> So here is my "major change in functionality":
> Go through the list of font names in "font" until you find a match
> (using the same rules that would be needed to implement the 10 August
> CSS proposal to read the prioritised list that follows
> "font-family"). Choose that font, with the weight, style, font size
> and leading specified in that line.
> Obviously, problems arise if and when Bembo is available, but not in
> demi-bold 12 pt. But these problems are identical to the ones that
> would arise from the following (which is ok according the the draft
> font-family: bembo new-century-schoolbook times
> font-size: 12pt
> font-leading: 14pt
> font-weight: demi-bold
> To summarise: I can see that Håkon's comment is a valid reason to
> entirely drop the prioritised font name lists from the draft
> specification. My proposal has exactly the same problems, but is more
> useful when and if the browser finds a match.
You're right with regard to the font-family: the styler will need to
determine if a given value can be successfully set or not. This is
unfortunate, and is only there because current window systems do not
provide adequate support for font selection: they are not able to
suggest the best alternative if the selected font is not
H1: font-family = helvetica sans-serif
Ideally, the font handler should know that helvetica is "sans-serif"
(just as it should know that red is #F00), but we can't deepend on
this yet. When this becomes possible there will be no need for a
prioritized list of font families.
Making other style properties depend on the outcome of this
unfortunate feature will make sure it remains and lessen the
motivation for more intelligent font handling.
Also, among the properties in the "set", font-family is among the
easiest to determine success or failure. Length units (e.g. font-size)
are harder since information about screen resolution/size often is
Bottom line: I'd rather drop the prioritized list of font-families
than institutionalizing it.
In any case, the specification should be clearer on these issues.
Hakon W Lie, W3C/INRIA, Sophia-Antipolis, France