There are bugs in the SPARQL specification with regards to EXISTS. The
RDF Data Shapes working group uses EXISTS, and other related mechanisms,
in SHACL [1].
W3C process for corrections is recognized generally to be inflexible. It
is normally to wait for the next WG to run and end which is a multiyear
cycle - that does not fit with the RDF Data Shapes WG timescale.
Community Groups can publish reports. These are not W3C standards. They
do provide a way to record consensus or enumerate alternatives. This
could be used to supplement the SPARQL errata process [2].
A suggestion is to use the W3C Community Group mechanism to describe a
solution to this specific area in a timely manner. The CG would document
a solution and create tests to pass over to the "RDF Tests" CG [3]. If
there is no single consensus on one solution within the SPARQL
community, including implementers and users, we can at least document a
small set of approaches and note the approaches taken by implementations.
Thoughts and comments?
Please indicate if you would join such an effort.
Andy
[1] http://w3c.github.io/data-shapes/shacl/
[2] https://www.w3.org/2013/sparql-errata
[3] https://www.w3.org/community/rdf-tests/
