- From: Brad Kemper <brkemper@comcast.net>
- Date: Tue, 15 Apr 2008 20:50:17 -0700
- To: Dave Crossland <dave@lab6.com>
- Cc: www-style@w3.org
On Apr 15, 2008, at 11:26 AM, Dave Crossland wrote: > > On 15/04/2008, Patrick Garies <pgaries@fastmail.us> wrote: >> Brad Kemper wrote: >> >>> I would prefer if it could be downloaded once and then used from any >>> page with an @font-face specifying the same font name, provided >>> there >>> could be some sort of quick check that it was the exact same font >>> (digital signatures or something, perhaps). >> >> I don't think that such a mechanism should be specified via CSS >> work. I >> really don't know how you'd specify such a thing anyway. > > Right; if the "url(font.foo)" is the same and already cached, it > should be loaded from cache. If the URL is different but happens to > have the same name or filename, it should be downloaded. > > Just like all other file formats. > > Why do people think font files are so special? :-) In terms of DRM, I was one of the first to say that they should not be treated differently. But in terms of scope, the specs for "font-family" and "@font-face" imply that "font-family" can use any font that is available, and that "@font" can make it available to the whole application. It does not say that the fonts available to "font-family" are limited to only those that were either already permanently installed on the computer or those that were provided by an @font-face on the same page. It does not say that "@font-face" only makes the font available to the current page or site. In fact, specifically mentions that the font should not be shared between applications, but it is silent about sharing between pages or sites. That implies to me that it is OK to keep the font resident within the application (the UA/browser software), and being resident, it is available to other pages (without an @font-face, but with the same "font-family", for instance). Also, the size of a typical font file would make me hesitant to include several across different subdomains of my site, even though I may really want to for the design. So I am very interested in ways that the probable negative bandwidth ramifications can be dealt with (or are being dealt with) by the implementors, and if there is anything we can do with the spec to help with that issue. > > > -- > Regards, > Dave >
Received on Wednesday, 16 April 2008 03:51:02 UTC