W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > November 2002

Re: docsync: RDF URI Reference

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Tue, 5 Nov 2002 09:46:39 +0100
To: <w3c-rdfcore-wg@w3.org>
Message-ID: <MABBLGKMPIJFCKFGDBEPGEELCBAA.jjc@hpl.hp.com>


I thought I might summarize the issue as I see it:

a: US ASCII or Unicode

We join a tradition of W3C specs that permit Unicode, possibly as original
character sequences in RFC 2396 terminology.

b: same doc references

In a web browser, as I understand it, a same doc ref "" or "#frag" should
not result in a retrieval action, when, for example the user clicks on it.
It is permitted, as I understand it, for the web browser to display an
absolute URI ref consisting of that same doc ref resolved against the in
scope base URI (possibly from an html base attribute).

There is strong feeling in some quarters that web browsers got confused,
with, I guess, at some stage, clicking on a "#frag" in a document with a
base URI other than the one it was retrieved from, caused a new retrieval
action (a bug).

In RDF we can pretend that our URI references are not retrieved, and so the
creation of an absolute URI ref that we do is analogous to the displayed
string case in the web browser - which would suggest it was cool to resolve
same doc refs against an xml:base.

However, if we consider Mike's DAML walker which allows you to click
through, I don't believe any of us would expect it to not do a retrieval
bacause some absolute URI started off life as an rdf:ID value with an
in-scope xml:base.

c: possible cause of action

Publish and be damned; do not play linguistic games to act as a figleaf.

When/If last call comment arrives ask for replacement text.

If the replacement text is adequate we go with it.

d: fall back position

Only permit xml:base in docs of mime type rdf+xml on the rdf:RDF element,
maintain the mapping to absolute URI and explicitly forbid Mike from doing a
retrieval for same doc refs. (I guess we could extend that prohibition to
other people too).



Jeremy
Received on Tuesday, 5 November 2002 03:42:42 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:53:58 EDT