Comments on CSS3 Fonts Module LC

In subclause 3.6: "font-size-adjust" - the specified value of 'auto' may be misleading.
The "font-size-adjust" property is intended to preserve the legibility of text rendered on various platforms (where a set of available fonts may vary) by adjusting the x-height  according to a defined aspect value, as illustrated by Fig. 19. The description of 'auto' seems to offer a promise of an easy way out where the aspect value is expected to be calculated by a UA for a given "first choice" font. However, it is not clear how this value would be calculated on a platform where the "first choice" font (e.g. "Verdana", as being used as an example in the spec) is not available. It seems that in that case the effect of the font-size-adjust would be 'no adjustment', which is contradictory to what the intended usage of the "font-size-adjust" should be.
Possible solution is to always require a specific 'number' be defined for an aspect value, effectively eliminating the 'auto' value and limiting the property value choices to either 'none' or 'number'.
[It may also be useful to mention that adjusting x-height of a font will affect ascender and descender values, which may results in collisions between them for a given font size / line height, or, depending on a UA rendering engine, it may possibly result in cropping of parts of the glyphs if the increase of ascender/descender values conflicts with a font bounding box.]

Subclause 4.1 and 4.3: Examples 8 & 9 (nitpick) - Considering that WOFF format is now a W3C Recommendation and is widely supported, would it be better to use '.woff' as a default example (here and elsewhere) of a downloadable font? The use of '.ttf' fonts as examples may be seen as recommended usage by some folks, I'd rather change that perception by using '.woff' for spec examples. The "Example 9" in subclause 4.3 deserves a special attention, where '.eot' is used as an example of source definition for use "with older, non-conformant user agents" while the use of '.ttf' implies conformance. [to what?]

Thank you,
Vlad

Received on Thursday, 15 August 2013 20:14:56 UTC