RE: Last call comments on WOFF

[Bert Bos:]
> UAs should not have to know via what protocol or from what IP address they
> retrieved the WOFF file. Indeed, the file may not have come from a single
> server (e.g., with bittorrent or Freenet) or not from any server at all
> (if it was "retrieved" from a URN). Specifically, that means that in the
> 4th paragraph of the Introduction all but the 1st sentence must be
> dropped, or replaced with something else. That something else could be,
> e.g., "UAs must implement the provisions of the xyz block."

There is no protocol, scheme or IP address restriction (origin != IP address). 
The restriction is that the origin of the font should match that of the
document using it. What scheme and port number are used, and what the host name
means for the scheme involved is completely irrelevant. (Although in practice
HTTP/HTTPS is going to be it 99.9% of the time). And as the format is primarily 
aimed as enabling web typography in web browsers - 'The primary purpose of the 
WOFF format..." - it's not clear why BitTorrent scenarios are a concern either. 
Any real world scenarios, pages or sites that show off the issue ? 

The WOFF intro paragraph then refers to origins and same-origin restriction as 
defined in HTML5, which is most definitely the primary expected context for WOFF 
use. 

Or are you disagreeing with HTML5 including these definitions and restrictions
as well ? If so then isn't this feedback for the HTML WG ?

Received on Wednesday, 26 January 2011 20:54:31 UTC