W3C home > Mailing lists > Public > www-style@w3.org > November 2008

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

From: Brad Kemper <brkemper.comcast@gmail.com>
Date: Wed, 12 Nov 2008 08:36:48 -0800
Cc: Håkon Wium Lie <howcome@opera.com>, "Dave Singer" <singer@apple.com>, <robert@ocallahan.org>, "Tab Atkins Jr." <jackalmage@gmail.com>, "Ian Hickson" <ian@hixie.ch>, <www-style@w3.org>
Message-Id: <321AF7D6-3CA0-4D86-A9DA-11E3B907099F@gmail.com>
To: "Levantovsky, Vladimir" <Vladimir.Levantovsky@MonotypeImaging.com>

On Nov 12, 2008, at 7:13 AM, Levantovsky, Vladimir wrote:

> One part of the compromise solution is using access control instead  
> of the root strings;

I like the idea of using access control to relax the "same origin"  
restrictions, so that I can link to a font from any of the style  
sheets on my site, not just the ones that happen to reside on the same  
server. In this case, I would be determining which servers were part  
of my site, and this would not be written into the font itself. I do  
not want to have to recompile or rewrap or re-compress the font every  
time I add a new server or contract with another company to serve some  
of my content (or their own content that has been customized for my  

In such a scenario, I think that a single bit on the font would be  
insufficient to communicate licensing rights. I would really need to  
know the following:

--Am I restricted from using the font on the Web at all?

--Can I use it only with a single server?

--Can I use it with all servers owned and/or operated by my company?

--Can I use it with all servers (or sub-domains of servers) which do  
business with my company, that are closely associated with my site,  
for which I have provided a custom style sheet to aid in it having a  
similar look and feel?

I personally would only use fonts that allowed the last option above,  
but I could see different foundries choosing one of the four options,  
and letting the market determine if their choice was acceptable enough  
to enough people.

On the other hand, how would they communicate this? By selling me all  
new "web-enabled" versions of the fonts I already paid for? I would  
resist such a proposal, but honestly, in the end I would probably bite  
the bullet and purchase them again if it were just a couple of fonts  
and it wasn't too expensive (type foundries, begin your drooling, over  
the possibility of extra cash). No one would like that much though.

BTW, once you are no longer relying on the browser to do your  
enforcement, you could still add watermarking to your compression  
scheme, in order to track down patent offenders. The UA doesn't need  
to pay attention to the watermark, but you could build a search engine  
to monitor watermarked fonts.

> the second part is font data compression that can also serve as  
> obfuscated font format. I believe we are in agreement that serving  
> compressed fonts on the web (and, thus, reducing bandwidth and  
> storage requirements) would be equally beneficial when using both  
> commercial and free fonts.

I don't think there is conceptually anything wrong with the idea of  
compressing the fonts beyond what is available via gzip. However, you  
also want it to be something that a browser can do and a desktop  
application can't, correct? This is where you run into a problem.  
There are other programs besides browsers that use the rendering  
engines of FireFox and Safari (gecko and webkit). This includes not  
only newsreaders and e-mail clients, but (with the advent of Google  
Docs and Google Gears), also includes word processors, spreadsheets,  
and presentation programs. As time goes on, we may see more and more  
of our non-Web-related desktop tasks use the rendering engines from  
Web browsers. Would you disallow them from using commercial fonts  
linked via style sheets? I don't think you would need to if the access  
control restrictions were in place, and if fonts pulled from the Web  
could not be simply dragged from a "temporary objects" folder into the  
permanent "fonts" folder of the OS.
Received on Wednesday, 12 November 2008 16:37:27 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:41 UTC