Hrvoje> Ben Wing <ben(a)666.com&gt; writes: Font-family is at best a hint, and sometimes completely misleading. See footnote 2. Hrvoje> I agree. The distinction between "face properties" and Hrvoje> "font properties" is often quite arbitrary. Not at all. A "face property" is "what we want to see", a logical property. A "font property" is "how we get what we want to see out of the system display mechanism." In the old X font model (pre-XLFD), there was only one font property AFAIK: its name. Even with XLFD, that's more or less true. Remember that originally a font was a collection of glyphs, divided into subsets called faces. Using multiple fonts was extremely expensive until computer typesetting became possible. Now the logical role of face is what we want to think about, and one of the implementation tools for differentiating faces is to change font. That is, the old relationship, face is a subset of font, has been inverted. This is the root of the "font model problem": different fonts, even from the same foundry, will have different models (in the sense that the units of "weight", "size", etc will be different and annoying or even hard to coordinate across fonts and font families, or even the font metrics will be different). But in modern applications, we need to do this kind of coordination, more or less automatically, on behalf of users who just want an 18-pt version of the 14-pt face they like a lot. I'm not at all sure I disagree with Kyle. It depends on how well we want to do things. I would be tempted to have a face model[1] with all the parameters we can imagine (size, slant, aspect ratio, family- list[2], color, strikeout, underline, etc) plus a series of plists named things like "x-font-hints", "windows-font-hints", etc. The semantics of the hints would be that they are derived from the face parameters and passed to XListFontsWithInfo(3x) and equivalents for other rendering engines. Then the information returned is compared to the face parameters to pick the best one. Footnotes: [1] I propose the terminology that "font models" are what we use to map the "face model" to implementation-dependent rendering facilities. I don't expect anybody to adopt it, since "font model" is embedded into people's brains, but I think it's worth proposing for the doubts it may stir up. [2] Families are not very useful in today's multiscript world, there are few Unicode fonts and few of even those are Uni-designed; rather they are glommed together from similar but not identical families--- this is apparently especially jarring in some TrueType Unicode fonts where an ISO-8859-1 font is used to display ASCII and Latin-1 and a visibly different font is used for Latin-2. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."