W3C home > Mailing lists > Public > www-webont-wg@w3.org > January 2004

Re: Gaspart's Comments

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Tue, 06 Jan 2004 12:53:50 +0000
Message-ID: <3FFAAFDE.6010106@hplb.hpl.hp.com>
To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Cc: www-webont-wg@w3.org

Draft Response follows.

I would not be unhappy to drop the PS which includes informative but 
non-consensus points. I think the reply is more helpful with them; I chose 
two points that are intended to balance out.

I am happy to send this, but suggest that Mike is probably the most 
appropriate responder.



Dear Prof. Gaspart,

thanks for your comments concerning the use of xml:base in the OWL Guide.

The Working Group has discussed these comments by e-mail [1].

The examples you highlight are not errors, and we do not intend to change them.

The use of xml:base in OWL is specified by RDF/XML Syntax (Revised) [2], we 
draw your attention to section 5.3.
specifically the words "any fragment part is ignored."

These words are further clarified in the test xmlbase/test013 in the RDF 
Test Cases [3].

The URL for the RDF/XML file in that test case is:


This shows an xml:base of "http://example.org/dir/file#frag",  two same 
document references rdf:about="" and rdf:ID="foo" and a relative URI 
rdf:resource="relpath". All three are resolved as if the xml:base were 
"http://example.org/dir/file". (This can be seen by examination of the 
corresponding ntriple file test013.nt).

Please reply, with a cc to public-webont-comments@w3.org, indicating 
whether you are satisfied with this response.

If you wish to follow up with comments on section 5.3 of RDF/XML Syntax or 
the RDF Test Cases, then such comments should be directed to 


on behalf of Web Ontology WG


[2]Beckett, RDF/XML Syntax Specification (Revised)

[3]Grant and Beckett, RDF Test Cases


Two comments of interest from the WG discussion are:
it is a bit weird that xml:base may be a  URI reference rather than simply 
Using a style of ontology definition designed to be less
error-prone, entity declarations and namespaces interact.  And the cleanest
set of declarations is

Entity decl   : <!ENTITY vin "http://www.example.org/wine#" >
Namespace decl: xmlns:vin = "&vin;"
Base decl     : xml:base = "&vin;"

Jeremy Carroll wrote:

> Noting
> http://lists.w3.org/Archives/Public/public-webont-comments/2003Dec/0031
> When I reviewed the guide example I was of the opinion that the xml:base value 
> was OK because RFC 2396 specifies resolving against base URIs, whereas XML 
> Base is explicitly permitted to have URI reference values. I've just looked 
> again at both documents and failed to find explicit text that justified the 
> behaviour of typical software which discards the fragment ID when asked to 
> resolve a relative URI against a base absolute uri reference.
> This suggests two possible responses and an action.
> Response 1) accept the comment and make the changes suggested (this avoids 
> failing to quote chapter and verse, but is a change)
> Response 2) a response something like (the **** are for my comment after)
> <<<
> The XML Base attribute can be used with a URI reference value:
> http://www.w3.org/TR/2001/REC-xmlbase-20010627/#syntax
> [[
> The attribute xml:base may be inserted in XML documents to specify a base URI 
> other than the base URI of the document or external entity. The value of this 
> attribute is interpreted as a URI Reference as defined in RFC 2396 [IETF RFC 
> 2396], after processing according to Section 3.1.
> ]]
> As indicated in this quote, and also in RFC 2396 section 5.1.1
> [[
> Within certain document media types, the base URI of the document can
>    be embedded within the content itself such that it can be readily
>    obtained by a parser. 
> ]]
> ****
> Thus when xml:base is used with a uri reference value, in fact only the URI 
> (without the fragment) is used as the document base.
> This is also clear in the algorithm in section 5.2 of RFC 2396 that makes no 
> reference to the fragment part of the base URI.
> ****
> Thus the examples selected have the exact same meaning as they would have if 
> the suggested changes were made. Given that ideally the documents do not 
> change between PR and REC we decline to make the change.
> comment: the **** part is dodgier than I would like.
> action 3) I could make a personal comment, or we could make a WG comment on 
> the XML Base spec noting that while it syntactically permits URI references, 
> it does not explicitly say "When the value of an xml:base attribute is a URI 
> reference, with a fragment, then the base URI is the corresponding URI, 
> without the fragment." and suggest that there should be an erratum to clairfy 
> this.
> If we make this as a WG comment then we could tell the commentator.
> ===
> I would be happy with (3) and either (1) or (2). I guess that if we decide for 
> (1) then the comment on XML base might be better as a personal comment, 
> whereas if we went for (2) then the comment might be a WG comment.
> Sorry for not having noticed this when I reviewed the examples.
> Jeremy
Received on Tuesday, 6 January 2004 07:54:03 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:56:56 UTC