- From: Chris Lilley <chris@w3.org>
- Date: Mon, 25 May 2009 18:56:21 +0200
- To: Zack Weinberg <zweinberg@mozilla.com>
- CC: Thomas Phinney <tphinney@cal.berkeley.edu>, W3C Emailing list for WWW Style <www-style@w3.org>
On Monday, May 18, 2009, 8:06:06 PM, Zack wrote: ZW> Thomas Phinney <tphinney@cal.berkeley.edu> wrote: >> > The CSS font-weight property does not take arbitrary numeric values, >> > only "nice even 100s". >> Why? That seems like yet another way of ensuring that different UAs >> get different results with the same CSS+fonts. ZW> My understanding is that CSS contemplates nine possible font weights, ZW> which happen to be assigned the ordered sequence of labels 100, ZW> 200, ... 900. That is correct, and at the time of writing it was asserted that fonts used only multiples of 100. However, in practice, they use other values as well. ZW> These don't (as far as I know) line up neatly with the ZW> weight properties in any font format, The 100..900 values in CSS2 were directly modelled on those in TT/OT. ZW> so there has to be some mapping ZW> established by the UA. There are guidelines for that mapping in the ZW> spec, but it may not be possible to nail things down perfectly. There are guidelines for the mapping in the CSS2 spec, which look well and good until a close reading shows that they are just an example and not actually testable. CSS3 Fonts responds to that by pulling even the suggested mapping. I don't think that is a good idea. It makes the spec less testable and implementations will have an undesired point of variation. I would prefer to see the mapping made normative. -- Chris Lilley mailto:chris@w3.org Technical Director, Interaction Domain W3C Graphics Activity Lead Co-Chair, W3C Hypertext CG
Received on Monday, 25 May 2009 16:56:50 UTC