RE: CSS3 @font-face / EOT Fonts - new compromise proposal

Hi David, 

> -----Original Message-----
> From: www-style-request@w3.org 
> [mailto:www-style-request@w3.org] On Behalf Of Dave Singer
> Sent: Monday, November 10, 2008 1:51 PM
> To: www-style@w3.org
> Subject: RE: CSS3 @font-face / EOT Fonts - new compromise proposal
> 
> 
> I'm going to try to summarize what I think I am hearing.  I 
> don't necessarily agree or disagree with what this, you 
> understand, I am trying to get clarity in at least my own mind.
> 
> On the 'serving' side, we are looking for an indication in 
> the font that shows whether it's freely usable or not.  The 
> 'allows embedding' 
> bit has been suggested, and that free fonts would have this 
> set and commercial fonts could have this clear, if they wish. 
>  This isn't obviously the right semantics, so that's question 
> one; is this the right indicator?
> 

Just to set the things straight - the presence of embedding bits in a
font indicates embedding restrictions. No restrictions (fsType = 0) mean
"installable embedding allowed". Therefore, free fonts would have fsType
field cleared and a commercial font will indicate whatever level of
restriction is specified by the font license.

> If the font indicates it's freely usable, then the serving 
> side MAY serve it as-is (but see below for recommendations).
> 
> If it indicates it's not freely available, the serving side 
> MUST 'obfuscate' the font in the chosen way, and it 
> can/should also use the access control methods from the W3C.
> 

So far so good.

> If a user-agent is requested to use an embedded font that is 
> not labelled as freely usable, and that font is not 
> 'obfuscated', the UA MUST refuse to use the font. 

If by UA you mean a browser - this can not be the case. UA doesn't know
up front whether a font is freely usable or not - it will/should use
whatever embedded font is served, whether obfuscated or not. I would say
that at this point UA should trust the serving side has done the
checking and made the right choice.
  
> The UA must also implement the access control restrictions, and 
> respect them if they are used (for anything, not just fonts). 
>  The UA MUST take care that the font is not generally 
> accessible to other applications while it's being used for 
> the web site it's embedded for.  The UA should exercise 
> reasonable care that it's not easy to find in its 
> non-obfuscated state.
> 
> The general font engines SHOULD NOT support the obfuscated 
> state directly;  the web UA should de-obfuscate before 
> passing it to the font engine.
> 
> For any font downloaded off the web, we recommend subsetting, 
> and compression.  Whether we need proprietary compression or 
> something like gzip is good enough would be the subject of 
> technical discussion.  We would recommend against using fonts 
> that disallow subsetting (though for the life of me I cannot 
> see why a font vendor would disallow it, or even why the 
> capability to indicate that is there).

I understand your premise and I agree with you on subsetting
restrictions, but I am not sure the recommendation against using full
font makes sense - what if I use a single font for all my web content
and there is simply no need to subset it. Why would we recommend not to
use the font [that doesn't allow subsetting] when it, in fact, should be
up to a content author to decide. What we probably should do is to
recommend subsetting whenever possible.

Regards,
Vlad

> --
> David Singer
> Multimedia Standards, Apple Inc.
> 
>

Received on Monday, 10 November 2008 23:05:35 UTC