Re: Specifying a Proxied URL on the CERN Server

Ari Luotonen (luotonen@neon.mcom.com)
Tue, 6 Dec 1994 20:55:55 +0100

Date: Tue, 6 Dec 1994 20:55:55 +0100
Message-Id: <199412061955.TAA03434@neon.mcom.com>
From: Ari Luotonen <luotonen@neon.mcom.com>
To: Multiple recipients of list <www-proxy@www0.cern.ch>
Subject: Re: Specifying a Proxied URL on the CERN Server

> Phil Hallam-Baker suggested doing an MD5 of the pathname to solve the cache
> name problem.

A very good suggestion. That's what the Netsite Proxy does, too.

> Our queries are likely to be short, but there are many terms and possible
> combinations
> in the worst case.
>
> Why not just apply the quotation stuff for URLs to the file names?

I'm not worried about the characters in the filename -- Unix
filesystem is quite liberal when it comes to that; I'm worried about
the length.

And BTW, the URL encoding is already used.

> I added 1.0 conditional gets based on if-moified-since and Expires headers
> on relevant outgoing requests (get, head, post).
>
> Will the cern server exploit this information effectively to minimize cache
> refreshes?

Yes.

> Is the CERN server going to do something reasonable with form submissions?

Yes; it forwards them untouched, and gives the results back untouched.

> We've got a fairly major collaborative WWW+email system rolling out.
> http://www4.ai.mit.edu/npr/user/root.html

That looks *facinating*!

-- Cheers, Ari --