- From: Andy Seaborne <andy.seaborne@topquadrant.com>
- Date: Wed, 13 Jul 2016 14:32:49 +0100
- To: public-sparql-exists@w3.org
On 12/07/16 23:08, james anderson wrote: > >> On 2016-07-12, at 23:28, Jeen Broekstra <jeen.broekstra@gmail.com >> <mailto:jeen.broekstra@gmail.com>> wrote: >> >> On 13/07/16 02:56, james anderson wrote: >> >> [snip] >> >>> yes, i did an initial example in the wiki, just to show what one >>> needs to be >>> able to look at in one plac just in order to comprehdne an issue. >> >> To be honest, I have no idea what I'm looking at there - assuming >> we're talking about the same wiki page: >> >> https://github.com/w3c/sparql-exists/wiki/Examples >> >> I see two SPARQL queries, the first followed by something which >> vaguely looks like a template for a query solution ( a lot of "s", >> "p" and "o" in various layers of bracketing) and then the sentence >> "the dataset as turtle", the second followed by "the intended >> solution as csv". >> >> Admittedly I haven't quite caught up with all the correspondence yet, >> but none of this really clarifies what the issue is, to me. >> >> Or do I perhaps misunderstand and is this merely a *template* you >> provided, that we collectively need to fill in with _actual_ examples? > > better, now you do understand. > it is a mock up which draws from two of the patel-schneider cases. > >> >> I'll be happy to assist with formatting if someone can provide some >> raw data: >> >> 1. a (minimal) sample dataset >> 2. a sample SPARQL query (we got those already) >> 3. expected outcome of query execution on the sample dataset >> >> I realize point 3 is exactly where the issue/dispute/uncertainty is, >> but picking _some_ example outcome gives us a good starting point, >> both for implementors to try their engine and see how it compares, >> and for us all to go back to the spec and see how we 'algebraically' >> arrive at that outcome. > > exactly. > the one purpose for that table is to get (at least) those three things > in one place on one page. > best would be if there were tooling which generated such a thing - in > a much more compact form than the pages accompanying the current test > suites, from the current test declarations. > > before embarking on a manual process, it would be good to > a) know that no tooling exists > b) agree on the layout. A small point to avoid confusion: the CSV format for SPARQL results is lossy. It uses only strings and numbers. The TSV form does put Turtle RDF terms as the cell contents. Andy > > best regards, from berlin, > --- > james anderson | james@dydra.com <mailto:james@dydra.com> | > http://dydra.com > > > > >
Received on Wednesday, 13 July 2016 13:33:22 UTC