I wrote > On Apr 23, 2008, at 3:03 PM, Jukka K. Korpela wrote: > >>> In a perfect world, 'Arial Black' shouldn't get bolded, as it >>> already >>> is a bold face with "weight=900". >> >> That would be the natural interpretation, but then "Arial Black" >> should >> not be a font-family value at all, just a particular font with >> font-family: Arial and font-weight: 900. However, many browsers >> probably >> treat it as a font family of its own. > > On OS X, 'Arial Black' is a _font-family_ with only one _face_ (with a > font-weight:900). > 'Arial' is a font-family with four faces : Regular, Italic, Bold and > Bold Italic (font-weight 400 and 700 resp). > > On Windows, it might be a particular _face_ within the Arial font- > family. I'm not close to the Windoze box to go checking.
Forgot to add a part: If it is a specific _face_ within a fon-tfamily, it shouldn't get bolded, as it alrady is a bold face. If it is a font-family, It could be argued that the browser might call its synthetic bolding algo's, but in this case, it shouldn't as the font-data already informs the browser that it is a bold face. Which is partly an answer to Jukka's: > Whether it's strictly a bug is a matter of definition. If "Arial > Black" > is taken as a font family of its own, with only one true weight > available, then it might arguably be acceptable to generated other > weights true algorithmic transformations. Actually that's what IE does > with Arial Unicode MS, for example; if specify font-weight: bold for > it, > IE will produce something bold, even though Arial Unicode MS has only > one weight available. This is comparable to "fake italics". If no bold is available, that is perfectly acceptable. Gecko, Opera and WebKit do the same. Philippe --- Philippe Wittenbergh http://l-c-n.com/ ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ List policies -- http://css-discuss.org/policies.html Supported by evolt.org -- http://www.evolt.org/help_support_evolt/