I agree with the sentiment, but not that ISPs represent consumers'
interests; generally, I find that they represent their own, with consumers'
as a distant second. Especially regarding caching.
I'm glad that people are talking about these issues; HTTP is AFAIK unique as
a transport protocol, in that it allows relaxation of semantic transparency
without explicit permission of the client or server. This will cause
problems if we just jump in (although some already are!)
Also, may be good to try and involve the W3C, as relaxing semantic
transparency potentially affects many things there (P3P, semantic web, etc)
On Thu, Nov 09, 2000 at 01:49:50PM -0800, Gary Tomlinson wrote:
At 11/9/00 1:32PM Hilarie Orman wrote:
I think this is more of an issue for Extensible Proxies than the
others.
I'm in complete agreement on the fault isolation issue. For other
things, while roles and responsibilities are ultimately important,
binding
agreements on paper are perfectly acceptable (i.e., they do not
have to have an on-the-wire format).
At November 9, 2000 1:37PM Keith Moore wrote:
I agree with this sentiment. I'm not trying to insist that the details
of such agreements be made a part of the protocol (though it would be
nice if we could do it, I suspect it would take a long time before we
could find a good way to encode the needs of real users into a protocol)
I think there still exists a need for the content distribution & delivery
providers to divulge their vested interests as well (paper is fine).
Traditionally, the CDNs we know today represent the content provider's
interests, while ISP's are likely to represent the consumer's interests.
Gary
--
Mark Nottingham, Research Scientist
Akamai Technologies (San Mateo, CA)