RE: .webfont Proposal 2

>-----Original Message-----
>From: [] On Behalf

>On idea 1, John, Sylvain and Håkon seem to have indicates that they
>could not accept this.

To make it clear: given our current EOT implementation, I certainly can't state I
couldn't accept rootstring matching :) And if I did, let me know when and I shall correct.

My understanding is that Mozilla and Opera are fully opposed to this. All the discussions I have
witnessed - in person, over email, publicly and privately - concur in this respect : any expectation
that browsers enforce same-origin restrictions embedded in the font file itself will not result in interoperable
implementations. It's a non-starter.

It is also my understanding, however, that it is OK for a font vendor to include URLs as part of their font's embedded
licensing info. But any such proposal must make it clear that user agents SHALL NOT use this information
for licensing enforcement purposes. It is very important to understand what this means: that a proposal that suggests a
browser MAY do this enforcement - i.e. it is optional and up the browser vendor - is also not going to be interoperable
in practice as some browsers will never do this.

This position, however, does not exclude using other mechanisms that do not depend on data embedded in the file e.g. Firefox 3.5
does not, by default, allow me to link to a font on John's site from my own web page. The built-in same-origin policy check
disallows this. John may, however, configure his server so as to allow me to do this. (If he has the license to do so,
of course). None of this is controlled by the font file itself.

Bottom line: a cross-browser solution cannot depend on user agents enforcing license rights. Note that this is not some kind of
endorsement or policy statement on my part; just a statement of my understanding of all the discussions I have followed thus far.

Received on Friday, 17 July 2009 21:20:59 UTC