- From: Erik Aronesty <earonesty@montgomery.com>
- Date: Fri, 2 Aug 1996 18:39:45 -0700
- To: "'http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com'" <http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com>
Is there a proposal to add globally unique id's to HTTP? There is a noticeable lack of unique object identifiers. Large objects are often duplicated on the Web....and UID's assigned to largish objects would reduce traffic. These UID's could comes through as response-header fields. "Content-Origin: url | some-other-distinguished-name" "Content-Identifier: md5-coded-hash | some-other-authenticatable-id" so when an indexer sees that....it can say "well...i already have that...so i'll store it as an alternate location of the same thing" or whatever.... ------------------------------------ or maybe even a little dumb authentication protocol: -client uses a HEAD method -server X's response contains the header: "Content-Identifier: purportedly-unique-content-id" -client already has a copy of this document given to him by Y -....asks if server X if Y's copy is to be trusted OR some such nonsense.........
Received on Friday, 2 August 1996 18:42:57 UTC