RE: Next step?

>OK.

>I tried to clarify this situation on this very list a while ago. All the font vendors who contributed, which wasn't many, >seemed to indicate that their licenses do/will require authors to impose same-origin restrictions. Earlier on in this >process, many font vendors expressed support for EOT and favourably mentioned its rootstring feature, which I also >interpreted as a desire for same-origin restrictions. Font vendors supporting WOFF are asking for a same-origin restriction >too.

>If in fact there are major font vendors who don't (or won't) require same-origin restrictions in their licenses, that would >make CWT much more useful. CWT's cause would be advanced if we could demonstrate that. Perhaps I can ask you if Microsoft >would permit use of its fonts on the Web without same-origin restrictions?

Yes, we do know that font vendors prefer same-origin enforcement. However, many have also come to understood that embedded domain labeling is not a practical mechanism for their users, just as they have come to understand that enforcing their own licenses is up to them. As a matter of fact, given that Ascender and others remain - despite your repeated objections and their presence on this list - quite interested in CWT, claiming this format to be useless for them is a straightforward contradiction of the available evidence. Asserting that this is the case based on license text you cannot yet produce is also unhelpful. I can only agree that this discussion should happen and that the technical alternatives are known and straightforward. 

Whether the outcome is 'pointless' to authors is your opinion and should ultimately be up to the latter. 

If you refer to the fonts that come with Windows, they are not licensed for web use. Same-origin policy or lack thereof has nothing to do with it.

>I think a putative Fonts WG should not discuss bare TTF/OTF fonts at all. It seems to me Microsoft has already ruled out >implementing them, so the only thing that would change that is market pressure, not a WG, so it would be pointless for the >WG to promote them as part of an interoperability solution (I don't think the "2 of 4" approach has value).

That was not the point. If both font vendor support *and* same-origin policy are essential to a successful web font format then two out of three of the current raw-font implementations are, by your own standard, even more useless than CWT. And will remain so whether Microsoft supports TTF or not. They should thus be so characterized and challenged.

>I also think a putative Fonts WG should not mandate a particular same-origin control mechanism. That would be >controversial, restrictive and unnecessary. What access control is required and how it should be enforced should be a >private contract between font vendors and Web authors. However, IMHO a Fonts WG should recommend that browsers have the >same "same-origin restriction by default plus CORS" behaviour as Firefox, because that gives authors a useful and >convenient tool to fulfill their font license obligations.

It certainly should; if lacking SOP makes a format useless then achieving interoperability requires same-origin policy by default whatever the on-disk format. If lack of SOP makes CWT pointless then surely WOFF would be equally pointless on those browsers that do not apply SOP to WOFF resources. Thus WOFF's interoperable success depends on SOP.
 
I agree that 'what access control is required and how it should be enforced should be a private contract between font vendors and web authors'. And that is exactly how I see the CWT SOP issue you keep raising. The statement applies just the same.

>I suspect Ascender and others have not fully appreciated that authors will have to implement Referer checking, nor how >problematic that will be.

Well, I can certainly try to chat with Ascender and Monotype one sixth or seventh time about it :)

>If authors are willing to step up and say "yes, we understand we have to implement Referer checking, and that's OK, it's >easier for us than the alternative of serving both WOFF and EOT Classic", then that would be good for CWT. If Ascender is >willing to step up and say "yes, it's OK to serve our fonts without imposing same-origin restrictions", that would also be >good for CWT.

As you said so well earlier, 'what access control mechanism is required and how it should be enforced should be private contract between font vendors and web authors'. In the absence of such contracts, all I have is the strong public and private statements of Ascender, Monotype and others asserting they want this to work.

>I've tried to elicit such statements in the past, without success.

Correction: you have been unsuccessful in getting a font vendor to agree that this issue made CWT pointless. Which should tell you something. Yet you keep bringing it up. (Sort of reminds me European referenda: we will keep polling you until we get enough Yes votes !)
 
>Your argument isn't quite valid, since we support non-Windows platforms where we don't deal with t2embed. 

t2embed is not the part that appends a header to the font data. My point is that the format itself is already known and handled internally. You're certainly not starting from scratch and even then it would remain a far smaller work item than implementing WOFF.

>But you're right that supporting CWT has a cheap implementation cost. Then again, there's a maintenance cost that lasts >forever, plus similar costs to all other browsers and other tools, plus a cost to authors of making format choice a bit >more difficult. Not to mention the cost of actually wrangling a CWT recommendation through a Fonts WG...

In my opinion, the cost of designing, building and supporting CSS3 Fonts, WOFF and the awesome new OpenType features Jonathan Kew is working on each dwarf the likely maintenance costs of yanking a known fixed binary header off a font file by several orders of magnitude. And I'm still probably being unkind of John and Jonathan.

And that's before we even factor in the likely benefits to authors which, even if marginal, add up as more and more of them get into web typography. But if you're willing to go out there and tell them all that supporting CWT is 'pointless' despite a) the stated willingness of font vendors to license their products in that format and b) CWT support on all IE browsers out there, all because of the maintenance costs of stripping ~100 bytes off a TTF, do go ahead. At least I wouldn't be the only one who's totally baffled...

Received on Friday, 23 October 2009 01:54:46 UTC