W3C home > Mailing lists > Public > www-international@w3.org > April to June 2011

font-size-adjust in CSS3 Fonts module

From: Peter Constable <petercon@microsoft.com>
Date: Mon, 2 May 2011 23:54:25 +0000
To: WWW International <www-international@w3.org>, "public-i18n-core@w3.org" <public-i18n-core@w3.org>
Message-ID: <71734097E7D39C439D341BF9F6C2457B3275F6C2@TK5EX14MBXC123.redmond.corp.microsoft.com>
Has anyone in the i18n WGs been reviewing work in the CSS3 Fonts module? In particular, I've recently come across the proposed font-size-adjust<http://www.w3.org/TR/css3-fonts/#font-size-adjust-prop> property and am concerned that it is poorly designed from an international perspective.

The intent is to allow adjustment of text size in font fallback situations. E.g., if I create a page set in Verdana, but also select Times as a fallback font in my CSS, then if text ends up being displayed in Times it will appear smaller than I intended because Times is drawn to different metrics than Verdana. The font-size-adjust property allows me a way to indicate a proportion to scale up the text in the event that it is displayed using Times rather than Verdana.

The problem from an international perspective is that the property is defined in a way that is intrinsically dependent on x-heights: the idea is to scale the text when displayed using Times (in my example above) so that the x-height is the same as what the x-height would be had it displayed using Verdana. It's bad enough that this design makes it hard to use even for Latin text. (The Web developer either has to have font-developer tools that will give her details about font-internal values-and then trust that the algorithm based on x-heights really does give the resulting typography I'd want, which isn't necessarily the case-or they have to use trial and error to figure out what values to use). But, from an international perspective, the property becomes meaningless for all those scripts (the majority) in which x-height isn't a relevant concept: in those cases, the only option is to use trial and error to figure out what values to use.

IMO, a much more sensible design for the scenario would have been a property that simply applies a scale factor to the font size: it makes it much easier to use even for Latin (I simply set some text using my preferred font and then set text with the fallback font, adjust the size of the latter to be the way I like, and take the ratio of the sizes), but it also makes it equally applicable to any script, not just Latin and a few others that happen to be like Latin.

Is there any review for things like this getting done by the i18n WG?



Peter
Received on Monday, 2 May 2011 23:54:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 2 May 2011 23:54:59 GMT