On Tue, 24 Jul 2007 14:50:34 +0100, Chris Lilley <chris@...157...> wrote:
(This is related to an implementation, not to what CSS can do).
TW> I'd really like to stress here that from a professional design point of TW> view, the CSS standard is totally inadequate for text designs.
Thats probably true, although the example you use to show it does not demonstrate that.
TW> There is TW> nothing remotely typograpically unusual about "Ultra Bold Condensed", TW> "Heavyface", "Extended #2", or any of the other dozen or so varients I see TW> every day which have no CSS equivalent.
Note that CSS can indeed describe a font which is ultrabold and condensed. The font-weight is probably 800 or 900 and the font-stretch is, well, condensed.
I really meant that the standard nomenclature of CSS is not expansive enough to be able to handle all the arbitrary names used in typography (old-style is an obvious ommision that's become very fashionable recently). It is certainly possible to set up a translation system from standard font naming to CSS naming/defining, but it is futile to think that this can be done for every case or that all the cases for which it can be done constitute a useful working sub-set for even quite ordinary typographical work. As such, the CSS representation of the fonts in a document must be regarded as one which will discard information and should not be the internal method of representing font variants.
To put it another way, if the names of a font and its variants are obtained from the OS then those are the names that need to be stored and used in the document unless platform portability is more important than having exactly the same font (although I can't imagine that happening very often); at that point a translation to CSS or any other standard can be attempted with the usual warnings about possible formating losses.
TW