Re: quick ping - ISSUE-104

I will be working in the coming days to pull together a consensus TAG 
position on issues relating to a Schema Datatype for CURIEs, or else to 
agree that the TAG has no such formal position.  Of course, TAG members 
such as Jonathan are welcome to pursue informal disucssions in parallel. I 
did want to clarify my understanding of the status of lexical and value 
spaces for anyURI in schema.  The intention in Schema 1.0 was to 
anticipate IRIs, which is why there is a reference to the XLink mappings. 
XSD 1.1 mentions IRI's explicitly.

Jonathan Rees writes:

> You can't support your assertion using the document you cite!  It says 
> quite clearly that the lexical and value spaces are arbitrary strings, 
> and that this mapping is the identity mapping:

Yes, for Schema 1.1, a bit less clear in Schema 1.0.  From Schema 1.0 [1]:

"The ·lexical space· of anyURI is finite-length character sequences which, 
when the algorithm defined in Section 5.4 of [XML Linking Language] is 
applied to them, result in strings which are legal URIs according to [RFC 
2396], as amended by [RFC 2732]. "  The use of the "algorithm" tends to 
ensure that internationalized forms are acceptable.  I don't think it's 
fair to say that this accepts any string, because of the appeal to RFC 
2396.  Then again, it's a little vague as to whether, for example, you are 
allowed to enforce scheme-specific rules.  Anyway, in the as yet 
unapporved Schema 1.1, the strings are indeed unconstrainted [2]:

"The ·lexical space· of anyURI is the set of possibly empty finite-length 
character sequences."

--and--

"This type should be used when the value fulfills the role of an IRI, as 
defined in [RFC 3987] or its successor(s) in the IETF Standards Track. "

So, Schema 1.0 at least tries to enforce some checking, but to my reading 
does it in a vague way.  In schema 1.1, the lexical and value spaces are 
arbitrary strings.   In both versions the mapping from lexical to value is 
the identity mapping.  In either version, it is a misuse of anyURI to 
apply it to lexical forms or values that don't conform to IRI, but that 
restriction is not completely enforced by the definition of the type in 
1.0, and is not enforced at all in 1.1.   Both versions accept forms that 
we would now recognize as IRIs, notwithstanding the fact that Schema 1.0 
was published before the IRI RFC.

Noah

[1] http://www.w3.org/TR/2004/PER-xmlschema-2-20040318/#anyURI
[2] http://www.w3.org/TR/xmlschema11-2/#anyURI

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








Jonathan Rees <jar@creativecommons.org>
08/31/2008 01:36 PM
 
        To:     Shane McCarron <shane@aptest.com>
        cc:     Ben Adida <ben@adida.net>, public-rdf-in-xhtml-tf@w3.org, 
Noah Mendelsohn <noah_mendelsohn@us.ibm.com>
        Subject:        Re: quick ping - ISSUE-104



On Aug 30, 2008, at 6:05 PM, Shane McCarron wrote:

>   The lexical space of anyURI is the complete collection of URIs as 
> defined in RFC 3986 (previously 2396 / 2732) [1].
> [1] http://www.w3.org/TR/2008/WD-xmlschema11-2-20080620/#anyURI

You can't support your assertion using the document you cite!  It says 
quite clearly that the lexical and value spaces are arbitrary strings, 
and that this mapping is the identity mapping:

                 The ·lexical space· of anyURI is the set of possibly 
empty finite- 
length character sequences.
                 The ·lexical mapping· for anyURI is the identity mapping.

Even if it did support what you are saying, this is just a working 
draft, and I don't think you can cite it in your document.

I'm not saying this is an easy rhetorical problem to fix - I'm just 
saying that it is consequential, and your obligation to those who will 
use your document in the future is to be careful and to dispel the 
smoke. You've disagreed with me repeatedly on both points, which 
leaves me uncertain as to how to proceed.

You need to live up to the rhetorical standard set by the document you 
cite [1]. Put the following, normatively, in Appendix A of the CURIE 
document:

                 The lexical space of CURIE is [fill in the blank, with 
citations and/ 
or internal references as needed].

                 The lexical mapping for CURIE is ...

                 The lexical space of URIorSafeCURIE is ...

                 The lexical mapping for URIorSafeCURIE is ...

Otherwise, when five years from now someone tries to figure out how to 
add CURIEs to SPARQL or OWL or RDF/XML or some other beast, they 
will ... shall I just say, be less than happy.

Jonathan

[1] http://www.w3.org/TR/2008/WD-xmlschema11-2-20080620/#anyURI

Received on Tuesday, 2 September 2008 14:43:00 UTC