Re: css3-fonts: should not dictate usage policy with respect to origin

On 28 Jun 2011, at 07:23, Glenn Adams wrote:
> It should be possible for these existing (non-HTML5) UAs to continue fetching BDF, OTF, TTF, etc.,

Of course those existing UAs can continue to do what they have always done...

> without becoming non-conforming under the auspices of a (finally, after ten+? years) complete CSS3-FONTS spec.

...but their behavior was never defined by a CSS3 Fonts spec, as that spec was no more than an experimental draft, as has been pointed out. If it is important to be able to claim "this UA conforms to <something>", then I suppose you can claim that it "conforms to the draft of CSS3 Fonts as of <whatever date>". Expecting that such a UA from the past will necessarily be able to claim conformance to the future final CSS3 Fonts recommendation is unreasonable, IMO.

CSS3 Fonts contains a *lot* of new material that was not present 10 years ago, as well as significant changes to previous behavior (e.g. the font matching algorithm). Do you expect the UAs in question are going to be updated to support all the rest of the current version of CSS3 Fonts? If not - and it seems unlikely to me - then they're necessarily going to be at best partial implementations, and will not be able to say "we conform to CSS3 Fonts" without extensive qualifications.

> Further, they should be able to support WOFF formatted wrappings of the same OTF/TTF fonts.

If the maintainers of these UAs wish to add such support, there is nothing to prevent it - a UA that is based on old standards, experimental drafts, and private vendor extensions and modifications can do anything it likes, including cherry-picking certain new features but omitting others. But it should not at the same time expect to claim conformance to published W3C recommendations.

JK

Received on Tuesday, 28 June 2011 09:31:49 UTC