- From: Government Linked Data Working Group Issue Tracker <sysbot+tracker@w3.org>
- Date: Thu, 18 Jul 2013 08:43:29 +0000
- To: public-gld-wg@w3.org
ISSUE-68 (QB validation gap): QB validation rules lack miss a potentially common case [Data Cube Vocabulary]
http://www.w3.org/2011/gld/track/issues/68
Raised by: Dave Reynolds
On product: Data Cube Vocabulary
An implementation report [http://lists.w3.org/Archives/Public/public-gld-comments/2013Jul/0008.html] has revealed that the validation rules in the spec are not as complete as we would have liked. They fail to detect a potentially common case.
The DataCube validation rules check that every Observation has a (unique) associated qb:DataSet (ic-1) and that every declared qb:DataSet has a structure definition (ic-2).
However, if the dataset associated with the observation does not have an explicit type declaration then ic-2 is not applied to it.
This is arguably an omission in the closure rules in phase one of the normalization algorithm [http://www.w3.org/TR/vocab-data-cube/#normalize-algorithm].
Should also have:
INSERT {
?ds rdf:type qb:DataSet .
} WHERE {
?o qb:dataSet ?ds .
};
The specification does say that implementations MAY use full RDFS inference which would cover this case. But it is only a MAY.
The hard question is what to do about this process-wise. Both the rules and normalization algorithm having be marked At Risk we can delete them but I don't think we can extend them without a reset to the process.
Received on Thursday, 18 July 2013 08:43:30 UTC