Re: New work on fonts at W3C

On Jun 25, 2009, at 10:28 AM, "Levantovsky, Vladimir" < 
 > wrote:

>> -----Original Message-----
>> From: Brad Kemper []
>> Sent: Thursday, June 25, 2009 1:08 PM
>> To: Levantovsky, Vladimir
>> Cc: Aryeh Gregor;; Jonathan Kew; www- 
>> Subject: Re: New work on fonts at W3C
>> On Jun 25, 2009, at 10:02 AM, Levantovsky, Vladimir wrote:
>>>> Referer-based methods are unreliable, and can block users of your
>> own
>>>> site as well as letting through users of other sites.  Some
> software
>>>> will strip referer headers, or even change them.  (For instance,
>>>> IIRC,
>>>> no browser sends Referer headers from an HTTPS site to an HTTP
>> site.)
>>>> Depending on referer checking isn't a good idea.
>>> Thank you. I am not a web developer so I would trust the experts on
>>> this.
>>> My point was that font vendors have so far proven to be reasonable
>>> people willing to cooperate, which is contrary to Rob's presumption
>>> that we would insist on a particular way of doing things (i.e. root
>>> strings).
>> So you would allow your fonts to be served in their raw format if
>> referrer based blocking was used? It could provide a large amount of
>> license violation, even if it wasn't perfect.
> This particular discussion had nothing to do with raw font format, it
> was relevant to discussing possible ways to subset EOT. And you lost  
> me
> completely when you mention "a large amount of license violation", Can
> you please elaborate further?


Sure. I believe one of your objections to using raw formats was that  
it would allow cross linking if other browsers did not enforce same- 
origin restrictions on them. I know you had other objections too, but  
on this point, given you link go other means of restricting cross- 
origin linking (such as referrer), would that satisy that particular  
objection to raw format distribution of fonts, if these other  
(admittedly imperfect) means were employed?

By "large amount of licence violation" (I meant to say "prevent" it,  
not "provide" it), I mean that if a Web site employed a mechanism that  
only let the font be served based on the referrer value being  
acceptable (i.e. only coming from licensed sites or staging servers or  
maybe even from Google Cache, etc.), that this would go a long way  
toward preventing mindless linking to fonts from people at other sites  
that did not understand the licence restrictions.

Received on Thursday, 25 June 2009 18:19:38 UTC