- From: Andy Seaborne <andy.seaborne@epimorphics.com>
- Date: Wed, 11 Apr 2012 09:09:36 +0100
- To: public-rdf-dawg@w3.org
On 10/04/12 20:34, Andy Seaborne wrote: > a > > On 10/04/12 17:01, Birte Glimm wrote: >> Hi all, >> >> since I was picked as the victim for handling AD-1 and AH-1, here's my >> attempt: >> http://www.w3.org/2009/sparql/wiki/CommentResponse:AH-1 >> http://www.w3.org/2009/sparql/wiki/CommentResponse:AD-1 >> >> In particular for AH-1 I would like some feedback for my arguments of >> why changing the BINDINGS semantics as suggested is problematic. > > In reverse order: > > ** http://www.w3.org/2009/sparql/wiki/CommentResponse:AH-1 > > +1 > > The technical reason is that what is requested is that > > BINDINGS ( ?x ) { ... } > BINDINGS ( ?y ) { ... } > > be the cross product (?x, ?y) but then what's > > BINDINGS ( ?x ) { ... } > BINDINGS ( ?x ) { ... } > > and generalise for more than one variable. Overnight thinking: If BINDINGS are allowed in groups, this just works out. The confusion point of BINDINGS ( ?x ) { ... } BINDINGS ( ?x ) { ... } is probably the empty (multi) set as I'd guess the normal writer is going to write two different (no value in common) sets. The join of different values for ?x being empty. BINDINGS ( ?x ?y ) { ... } BINDINGS ( ?x ?z ) { ... } is chaotic but explainable, maybe even sensible. Andy
Received on Wednesday, 11 April 2012 08:10:08 UTC