- From: Gregory Williams <greg@evilfunhouse.com>
- Date: Sat, 11 Feb 2012 14:31:19 -0500
- To: Axel Polleres <axel.polleres@deri.org>
- Cc: "Lee Feigenbaum" <lee@thefigtrees.net>, "SPARQL Working Group" <public-rdf-dawg@w3.org>
On Feb 11, 2012, at 2:22 PM, Axel Polleres wrote: >> Why do we need to define the equivalence or potential optimization spots at all? > > If there had been an easy way to define an equivalence with what we have, it would probably have simplified answering/addressing > the comment directly. > > [...] > >> Why can't we leave other semantics and optimizations to implementors and any future WG to nail down? > > Sure, I was just summarizing what I had understood as the two options to address the comment from the last Telco. > Just leaving things as they are, and pointing to a future WG is another Option. Should we call it Option3? I wouldn't mind having some text that raises this issue in the document, and suggests that implementations may seek to optimize cases where DISTINCT is used. I just don't see why we need to fully define an equivalence with one of many possible other path semantics. This feels similar to the slight hand waving we did with the overlapping semantics of MINUS and NOT EXISTS (we notably stopped short of providing a full definition of where the two negation styles differ). .greg
Received on Saturday, 11 February 2012 19:31:45 UTC