[sc34wg3] TMQL: Atoms "iri"

[Lars Heuer]

Comments against TMQL draft dtd. 2008-07-15.

4.2 Atoms
=========

[10] iri
--------
That production is ambiguous.

Example:
a.
    %prefix mailto http://psi.example.com/
    
    "mailto:someone at example.org"
    
b.
    "mailto:someone at example.org"


How is "mailto:someone at example.org" interpreted? For (a) it would be
interpreted as QName and for (b) it would be an IRI?

In my opinion it makes sense to change / restrict the "iri" production
to::

    iri ::= " IRI "

But even if we restrict the production to my proposal, the following
string has to be interpreted as IRI, because of production ``[13]
IRI``::

    "this-is-a-string-that-does-not-contain-invalid-iri-chars"

Acc. to production [13] nobody can decide if that construct is meant
as IRI or as string.

Example:

    "mailto:x at abc.org" << atomify 
    
Does that mean, that 
a) all occurrences with the value "mailto:x at abc.org" and datatype
   ``xsd:anyURI`` should be produced?
b) all names and occurrences with the value "mailto:x at abc.org" and
   datatype ``xsd:string`` should be produced?

In conclusion, the solution to embedd IRIs into quotes ``"`` seems not
to solve the issue "Grammar Rule for IRI inconsistent" (c.f.
<http://www.itscj.ipsj.or.jp/sc34/open/1046.htm#id2478390>).

If TMQL would drop one string notation (i.e. '([^']|\')*' which is
also unsupported by CTM) it could use that notation for IRIs.

Oh well, now you've read this looong mail and the solution is that
simple that it fits in one sentence. ;)

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>