- From: Henrik Frystyk Nielsen <frystyk@microsoft.com>
- Date: Tue, 18 Jul 2000 20:29:03 -0700
- To: "Miles Sabin" <msabin@cromwellmedia.co.uk>, <xml-uri@w3.org>
> The Http Extensions Framework (RFC 2774) has an identification
> mechanism similar to XML Namespaces. It also uses URIs as
> identifiers and might also be vulnerable to some of the same
> issues. It specifies absolute URIs only (so it's avoided one
> of our problems), but, unfortunately, it hasn't specified any
> URI comparison mechanism, so the canonicalization issues
> remain.
>
> See,
>
> ftp://ftp.isi.edu/in-notes/rfc2774.txt
You are right that it is very similar in intent but as you say, HTTP-EF
doesn't allow relative URIs and it is certainly the intent that it uses
the comparison algorithm described in the HTTP/1.1 spec, see
http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.2.3
The reason why it here makes sense to not allow relative URIs is that all
URIs exported by a Web server by definition are absolute - even in the
weird case of the "*" request-URI. There is therefore always an explicit
base URI which both parties (client and server) in the communication know.
Henrik Frystyk Nielsen
mailto:frystyk@microsoft.com
Received on Tuesday, 18 July 2000 23:30:01 UTC