Re: New work on fonts at W3C

>> You don't need a new format to do these two things, you can do this  
>> by tweaking the contents of the name table in TrueType/OpenType fonts:
>> 1) Make the family name "No Tresspassing" and the style name " for  
>> web use only" in for all web fonts.  Better yet, put in a GUID  
>> string based on the purchaser/site.
>> 2) Change the contents of the license record to say "This font  
>> licensed to xxx by yyy for use on site zzz.  All other use  
>> restricted and governed by the terms below.  For more information on  
>> this excellent font please visit"
>> 3) Include a sample @font-face definition in a text file that  
>> defines the set of font faces and their associated style attributes  
>> (i.e. not the obfuscated style names above).
>> That seems to satisfy your requirements; "normal" use in desktop  
>> applications will not be possible and the font data will be clearly  
>> marked as being associated with a given site in the license metadata.
> These sound like really great ideas. I wonder if one of the fields
> (such as the license record) could contain a machine-readable list of
> sites that the font was licensed for, and if that list could be used
> for CORS (for an HTTP server serving the font to read and include in
> the header, or for the browser to read directly as though it was a
> header).

Things like machine-readable lists of domains are equivalent to a root
string solution, like EOT, and suffer the same problems: moving the
font file around between different servers (e.g. stage to production)
requires constant updating of the font data, is difficult to work with
locally (e.g. drive/path inclusions in EOT spec(!)), and breaks down
completely when used with caching servers (e.g. Akamai).

That said, I think it would be interesting if browser vendors could
agree on ways of exposing license information to users examining a given
page's design.  So an author who wonders "Hmm, what font is this page
using?" could easily find out what the font is, who made it, what the
license is and where to find out more.  David Berlow's PERM table
idea[1] is interesting in this respect, it seems like there might be
finer-grained information in there to capture for example whether a font was
generally available or a font made for custom-use only (e.g. a "Time
magazine font" intended only for use with Time's site).  But I still
think information like this should be informative rather than used for
some sort of machine-readable license handling scheme. 

John Daggett
Mozilla Japan


Received on Friday, 19 June 2009 03:06:29 UTC