Re: Proposal: Vary on Cookies

Richard Vermillion:
>
>Dave,
>
>Thanks for your response.  I did take a look at the Transparent
>Content Negotiation draft, but it didn't seem to address my problem
>directly.

Yes, TCN does not address this problem directly.  As you correctly
point out, you can do some things with special remote variant selection
algorithms which take cookies into account, but I see this as a
technology option for the far future.

In my opinion, it is too early to consider extending Vary for matching
on parts of headers.  We do not even know yet whether HTTP/1.1 proxies
will implement all the extra optimisations made possible by plain
HTTP/1.1 Vary.

Also, just doing Etag based conditional GET processing for your
negotiated content will give a very nice improvement over HTTP/1.0
already.

>-rv

Koen.

Received on Friday, 13 June 1997 11:41:29 UTC