- From: Dave Singer <singer@apple.com>
- Date: Wed, 05 Nov 2008 15:04:32 -0800
- To: Håkon Wium Lie <howcome@opera.com>, Bert Bos <bert@w3.org>
- Cc: "www-style@w3.org" <www-style@w3.org>
At 23:52 +0100 5/11/08, Håkon Wium Lie wrote: > - to add a new table with root strings, i.e, a list of sites and > pages that are allowed to use the font > >I've done some soul-searching on this and reached the conclusion that >I can live with the first part, but not the second; I don't want web >resources to have inherent page/site restrictions. It breaks with a >fundamental principle of the web, that web resources should be >reusable. Further, it may violate the "fair use" doctrine of copyright >law. If root strings are accepted for fonts, content providers would >soon ask for root strings in images and video as well. Also, it's not clear if they'd 'work' correctly. Here are two simple examples: 1) When we post new content to our web site, it's first 'staged' on internal staging servers, and tested. It is then moved into production. This process relies on URL-neutrality, particularly using relative URLs intelligently. The fonts would be served from different servers (staging and production) and the root strings might test OK and then fail in production etc. 2) When our content is posted, we selectively 'akamai-ze' it. That is, some content is allowed to flow into the akamai cloud, and the base URLs that are published for it now point at akamai. Akamai, in turn, knows how to get back to us to find the base data that they are providing a global distributed cache for. The Akamai URLs and the base URLs are not the same. I am sure that there are many other problems here, at least around caching and so on. Root strings seem to be rather like a 'come from' statement [<http://www.fortran.com/come_from.html>], (or even a 'dont come from', reference needed), and analysis has shown that these can be confusing in use... -- David Singer Multimedia Standards, Apple Inc.
Received on Wednesday, 5 November 2008 23:05:53 UTC