Create 20190410-minutes.md

meetings/2019/20190410-minutes.md

@@ -0,0 +1,62 @@
+# ShEx Telecons
+
+## Telecon 2019-04-10  Minutes
+
+* Participants: Dimitris, Tom, Nishad, EricP, Kat, Jeremie
+* Regrets: Harold
+
+Call in Zoom # https://zoom.us/j/441496948
+
+* Chair: Dimitris
+* Scribe: Tom
+
+# Address PPS comments on Shape Map in ShEx 2.1
+
+ericP: the shape-map spec should in no way be in the critical path for evaluating the semantics spec, but we may as well get this out of way.  ShEx presumes a set of node-shape pairs.  How you get to that set is orthogonal to ShEx semantics per se.  Shape maps should not be in the critical path to reviewing semantics.
+
+**PROPOSED**: the note in [ShapeMap usage](https://shexspec.github.io/shape-map/#usage) addresses [pfps's issue](https://lists.w3.org/Archives/Public/public-shex/2019Mar/0006.html) "`It is still the case that the map from query ShapeMaps to fixed ShapeMaps might fail.`"
+
+ericP: +1
+TomB: +1
+Dimitris: +1
+Jérémie: +1
+Kat: +1
+Nishad: +1
+
+**RESOLVED**: the note in [ShapeMap usage](https://shexspec.github.io/shape-map/#usage) addresses [pfps's issue](https://lists.w3.org/Archives/Public/public-shex/2019Mar/0006.html) "`It is still the case that the map from query ShapeMaps to fixed ShapeMaps might fail.`"
+
+Regarding pfps' comment `It should be stated somewhere that ShapeMaps can contain actual RDF nodes, which can be blank nodes (and not blank node IDs).  What is the syntax for this?  It doesn't show up in the ShapeMap grammar.`
+
+ericP: If we use SPARQL-matching semantics, with a query pattern with a BNODE, e.g. ```{FOCUS <p> _:abcd}```, `_:abcd` will be treated as a variable and will match anything in the graph. But if we use not-standardized but omnipresent triplesMatching function in every graph store, `_:abcd` would match only that BNODE in the graph.  SPARQL semantics are more inclusive; triplesMatching more restrictive.  What we get is a bit more control.  So far, in the ShEx spec: has gone beyond SPARQL because can look into the lexical form of BNode.   In most RDF parsers, must set special flag.  Original concept of blank nodes: referential integrity _within_ a graph.  But then in SPARQL 1.0, could use variables to match blank nodes.  SPARQL forced the hand of RDF by clarifying that bnodes are scoped to dataset, not graph.  Would be less controversial to just use SPARQL, but also less powerful.
+
+Dimitris: Let's go with triplesMatching, as Jeremie also prefers.  We should get approval from other implementers.
+
+ericP: I know this isn't implemented everywhere: rdflib makes it difficult, and Greg does not like.
+
+
+**PROPOSE**: https://github.com/shexSpec/shape-map/pull/17 addresses pfps's question about SPARQL-matching.
+Votes are on the PR.
+
+## QueryMap
+```
+{FOCUS <p> _:abcd}
+```
+## Graph
+```
+<s1> <p> "hi" .
+<s2> <p> _:abcd .
+<s3> <p> _:efgh .
+```
+SPARQL semantics gives us (`<s1> <s2> <s3>`) while triplesMatching gives us only (`<s2>`)
+
+ACTION: Eric to explain re: SPARQL-matching vs triplesMatching.
+
+# Community updates
+
+Tom: Hack Day at DC-2019 in Seoul, September, will focus on application profiles (and ShEx).
+
+Tom: https://ericprud.github.io/shex-form/ - implements an oft-requested feature.  Generates forms: either blank forms, or forms populated with data.
+
+# EXTENDS
+
+ericP: Last week, Jeremie concerned that complexity of error reporting.  But what if you try to address _without_ EXTENDS.  Also: Harold had raised use case about discriminating types.