W3C home > Mailing lists > Public > www-rdf-comments@w3.org > April to June 2003

Re: xml:base irrelevant to same document reference

From: Graham Klyne <gk@ninebynine.org>
Date: Mon, 16 Jun 2003 08:30:59 +0100
Message-Id: <>
To: James Clark <jjc@jclark.com>
Cc: www-rdf-comments@w3.org

Yes, it's awkward.

Sidestepping your main point, I'll note that rdf:nodeID isn't intended to 
be like rdf:ID, and I think should be treated separately.  In particular, 
there is no transformational relationship between a nodeId value and any 
URI -- it is purely a local tag (bound identifier) within an RDF graph to 
allow representation of some graphs that would otherwise be unserializable.

Concerning same-document references, I think the original text pre-dated 
xml:base, and I've heard Roy Fielding say it was intended to address the 
point that, in "normal web use", following a fragment identifier should not 
cause multiple retrieval of the same document.  I've heard no indication 
that "escaping" a feature like xml:base was explicitly part of the 
intent.  I'm not sure if, given how things have evolved, there's any 
entirely consistent treatment of xml:base (independently of RDF concerns).


At 07:07 16/06/03 +0700, you wrote:
>Graham Klyne wrote:
>>The problematic "same document reference" wording from is proposed to be 
>>revised in the forthcoming revision of RFC2396 [1] [2].  With this 
>>change, I think that using xml:base is entirely consistent with RFC2396.
>Thanks for the pointers.  I see you have been through this issue 
>exhaustively, but I have to say I am still left feeling a bit uneasy. The 
>revised RFC2396 version still says "When a same-document reference is 
>dereferenced for the purpose of a retrieval action, the target of that 
>reference is defined to be within that current document or message; the 
>dereference should not result in a new retrieval."  Consider the following 
><p xml:base="http://www.example.org/foo"><a name="xyzzy"/></p>
><p xml:base="http://www.example.org/bar"><a href="#xyzzy"/></p>
>Even with the proposed RFC2396 revision, the second href will refer to the 
>first href (at least I think so), whereas in the analagous case in RDF it 
>will not.  Surely there is something wrong somewhere if XHTML and RDF 
>behave inconsistently in this respect. And I have never come across 
>another XML vocabulary that scopes IDs to the current base URI.

Graham Klyne
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E
Received on Monday, 16 June 2003 03:35:41 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:15:20 UTC