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

Yes, I can confidently assert that interop is improved when all UAs produce the same output given the same input. If that’s controversial there is no point in discussing further. If we agree that same-origin should be the default for font requests from @font-face with an optional opt-out then this should be the normative behavior. And since this normative behavior has a direct impact on @font-face processing by UAs it is best required  by css3-fonts.

So there are two questions here that should remain separate: a) what the default origin policy should be for the src descriptor of @font-face b) where it should be defined. I’m only discussing b) at the moment. If we want to argue a), let’s start a separate thread.

Improving on the situation: vs. what we have today whereby some browsers will successfully load a given font while others will not. Again, assuming the agreed default should be same-origin, please explain how requiring this behavior in a separate document is any better and/or more likely to yield introp than specifying it in the document that defines the exact feature it is specifically tied to. This is conceptually no different than HTML5 defining the same-origin tainting mechanism for canvas elements.

A situation where a browser can pass the entire css3-fonts  test suite but remain incompatible with the way other browsers process the exact same feature is a botched standardization job imo. Theoretical perfection held up as an enemy to the practical good.

(Again, I don’t suggest defining the actual header, values and other protocol details inside css3-fonts; the latter would just require same-origin for src and refer to the spec that defines the mechanics).

From: Glenn Adams [mailto:glenn@skynav.com]
Sent: Wednesday, July 20, 2011 6:04 PM
To: Sylvain Galineau
Cc: Christoph Päper; W3C Style
Subject: Re: css3-fonts: should not dictate usage policy with respect to origin

Well if you like to go asserting, then I shall assert as well: the burden of proof should be on you that SOR is required to obtain interoperability for CSS3 Fonts functionality; if you can demonstrate that SOR functionality pertains to CSS3 Fonts functionality, then I will gladly withdraw my assertion.

You say "improve the situation". What situation?

On Wed, Jul 20, 2011 at 6:04 PM, Sylvain Galineau <sylvaing@microsoft.com<mailto:sylvaing@microsoft.com>> wrote:
Two browsers implement SOR interoperably. Two others do not. What other evidence is needed ? The burden is on you to prove that specifying it in another document without any mention of it in CSS3 Fonts is more likely to improve this situation. The implicit claim that more specs results in better interop seems dubious to me, at the least.

From: www-style-request@w3.org<mailto:www-style-request@w3.org> [mailto:www-style-request@w3.org<mailto:www-style-request@w3.org>] On Behalf Of Glenn Adams
Sent: Wednesday, July 20, 2011 11:32 AM
To: Christoph Päper
Cc: W3C Style

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

I would like to see evidence of how specifying or not specifying SOR aids or detracts from "CSS3 Fonts interop". I have seen no evidence to date. The functionality of CSS3 Fonts is unaffected by, and entirely orthogonal to SOR or fetch/access algorithms.

G.
On Wed, Jul 20, 2011 at 12:02 PM, Christoph Päper <christoph.paeper@crissov.de<mailto:christoph.paeper@crissov.de>> wrote:
Sylvain Galineau:
> The right place to define requirements needed to achieve CSS3 Fonts interop is the CSS3 Fonts spec.
Sure.