On 10/08/11 03:01, Alexandru Todor wrote:
> While running the latest version of the extraction framework on the
> German data dump, I got some property names that have round brackets in
> them "(" ")" like for example:
> "http://de.dbpedia.org/property/Austragungsort(e)" . The problem is that
> Pubby crashes when a page is requested that contains such a property.
> The crash is due to Jena, which executes a remote sparql query on
> Virtuoso and recieves invalid XML as a response. The problem is that I
> don't even know where to fix the bug, The URL RFC [1] Section 2.2 states
> that round brackets can be used without escaping them, the URI RFC [2]
> section 2.4.3 also doesn't mention them being dissalowed so the
> extracted URIs should be valid. However I don't know if the RDF spec
> allows property names to contain round brackets .
The RDF spec does allow that, but the problem is that there is no way to
serialize such property URIs in RDF/XML. Specifically the fact that your
property URI ends with a closing bracket is a problem.
It's a known issue that there are valid RDF graphs that can not be
represented in RDF/XML (which is one of many good reasons not to use the
RDF/XML syntax format).
> Is the extracted data invalid, or is there a rdf-spec problem ?
The extracted data is invalid XML, yes: an XML element QName can not
contain brackets. Unfortunately, there is no right way to do this in
your case.
The RDF/XML spec recommends that a writer tries to split the URIref
after the last non-NCName character and use an ad-hoc namespace
declaration, but if the last character of the URIref is a non-NCName
char (such as the closing bracket in your property URI), there is no way
to split it, and a writer tool should report an error (apparently
Virtuoso has opted for not giving an error but producing invalid XML
instead - either way the communication breaks down).
The only reliable way around the problem is to use a serialization
format that does cope with all legal RDF properly, such as N-Triples or
Turtle.
Cheers,
Jeen
------------------------------------------------------------------------------
Get a FREE DOWNLOAD! and learn more about uberSVN rich system,
user administration capabilities and model configuration. Take
the hassle out of deploying and managing Subversion and the
tools developers use with it.
http://p.sf.net/sfu/wandisco-dev2dev
_______________________________________________
Dbpedia-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion
