W3C home > Mailing lists > Public > www-font@w3.org > July to September 2009

Re: A way forward

From: John Daggett <jdaggett@mozilla.com>
Date: Fri, 24 Jul 2009 01:56:02 -0700 (PDT)
To: Sylvain Galineau <sylvaing@microsoft.com>
Cc: www-font <www-font@w3.org>
Message-ID: <15794628.41891248425762114.JavaMail.root@cm-mail02.mozilla.org>
Sylvain Galineau wrote:

> >If all browser vendors are going to make the effort to support a
> >new format it should be one that behaves consistently across
> >browsers and supports all forms of OpenType fonts.  Supporting some
> >form of EOT produces inconsistent behavior, it limits what can go
> >into the standard in order to fit the behavior of older versions of
> >IE.
> 
> Wait. Where did anyone say you had to conform with IE's legacy bugs
> and limitations ? Who said the standard had to conform to legacy IE
> ? Accepting font files with an EOT header does not mean acting like
> IE6. We're the ones who will have to bend over backwards to plug the
> gap around @font-face, file types etc. Like we did with CSS2.1.
> That's our burden, not yours.
> 
> If the standard ends up saying IE is missing features, we're the ones
> who have to catch up. Surely you don't object to that ?
> 
> In the meantime, there would be a useful amount of cross-browser
> interop for TTFs. I'd rather have that than nothing at all. So do a
> lot of people, apparently.

If we're talking about supporting a common format, that format should
have a defined behavior.  If we use a legacy format like EOT, then the
"new" behavior gets mixed up with the old behavior; is IE supposed to
reject cross-site EOT loads?  Or is that just a deficiency of EOT in
general? If it's clear that EOT is a Microsoft-only format, those
problems remain squarely on your shoulders, we don't have to share
those associations.

Frankly, the whole legacy issue seems like bull to me.  It's a very
minor piece of code to put in a shim layer in IE that swizzles
whatever commonly agreed upon format folks come up with (.webfont,
ZOT) into an EOT format.  Since you're part of the OS, you can push it
out with a minor OS update and worry about fixing the overall
@font-face behavior later.  Maybe even fix CFF support while you're at
it.  Would you like me to write it for you?  Or just feed Sergey
a few beverages of his choosing?

> Incidentally, it would be very helpful if you could summarize all
> your objections in one place, so it doesn't feel like a drip-feed of
> the next issue as soon as the last obstacles are removed.

Um, that seems unproductive at this point.  Microsoft has been pushing
EOT for well over a year now and no other browser vendor has shown any
interest in it.  Berating me isn't going to change that.

> >This is just what Jonathan Kew pointed out [1].  For example, older
> >versions of IE don't support same-origin restrictions so cross-site
> >font linking would work with these versions but wouldn't work with
> >browsers supporting same-origin restrictions.
> 
> You're absolutely correct. But whether that is acceptable is really
> up to font vendors. WebKit does not implement same-origin
> restrictions either afaik yet that did not stop you from being
> incompatible with them so I don't see why the same issue should
> suddenly be this great hurdle that puts a burden on everyone as soon
> as IE is involved. This is an issue we all have to find agreement on
> today. Laying it at the feet of legacy IE releases may be very
> convenient, but it's totally unhelpful imo.

If we're going to create a new format we should be creating one with
consistent, agreed-upon behavior, not one that simply creates a new
set of inconsistencies.

> I'm glad we agree it's not just about browser vendors. So you would join a
> Fonts WG provides :
>
> 1. It standardizes .webfont or ZOT; no rootstrings;
> 2. Other browser vendors are involved ? Which ? Microsoft ? Opera ? Apple ? All ?
> 3. A plurality of font vendors are comfortable with the proposal. We agree on that.
> 
> Is that fair ?

Sounds reasonable.  I think an informal group would be better to start
to iron out a rough set of details, that would be faster and simpler. 
Then charter a WG as needed.  It would be best to have font vendors,
both large and small.
Received on Friday, 24 July 2009 08:56:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 11 June 2011 00:14:03 GMT