See also: IRC log
<scribe> scribenick: dimitris
<Arnaud> PROPOSED: Approve minutes of the 28 April 2016 Telecon: http://www.w3.org/2016/04/28-shapes-minutes.html
arnaud: let's start, propose to approve minutes from last week
RESOLUTION: Approve minutes of the 28 April 2016 Telecon: http://www.w3.org/2016/04/28-shapes-minutes.html
<Arnaud> PROPOSED: Open ISSUE-158
<pfps> ISSUE-158
<trackbot> ISSUE-158 -- ill-typed literals do not always trigger a validation result -- raised
<trackbot> http://www.w3.org/2014/data-shapes/track/issues/158
<Arnaud> issue-158
<trackbot> issue-158 -- ill-typed literals do not always trigger a validation result -- raised
<trackbot> http://www.w3.org/2014/data-shapes/track/issues/158
<TallTed> +1
+1
<pfps> fine by me to open 158
<hknublau> +1
<kcoyle> +1
<pfps> +1
RESOLUTION: Open ISSUE-158
<trackbot> issue-123 -- Shall we unify the syntax of sh:directType and sh:class? -- open
<trackbot> http://www.w3.org/2014/data-shapes/track/issues/123
<Arnaud> PROPOSED: Close ISSUE-123, dropping sh:directType
arnaud: we talked about this, it seems there is consensus to drop sh:directType
<hknublau> +1
<pfps> +1
<jamsden> +1
+1
<marqh> +1
RESOLUTION: Close ISSUE-123, dropping sh:directType
<trackbot> issue-135 -- Should sh:and/sh:or/sh:not/sh:valueShape support constraints too? -- open
<trackbot> http://www.w3.org/2014/data-shapes/track/issues/135
<pfps> there have been several emails
arnaud: we talked about issue 135, people liked the idea for simplification and Holger sent an email with a proposal
hknublau: i noticed I made a mistake with the original proposal, I was propagating property constraints, the solution is simpler without changes but I am not ready for a complete proposal
<Arnaud> https://www.w3.org/2014/data-shapes/wiki/Proposals#ISSUE-135:_and.2For_syntactic_sugar
pfps: there is an nice proposal from Eric, if extended properly it could be a good candidate
arnaud: ericP is not here and cannot answer questions, maybe best is to give WG another week
<kcoyle> second set of comments from
<kcoyle> Baker https://lists.w3.org/Archives/Public/public-rdf-shapes/2016May/0004.html
arnaud: there were some comments on the editors draft from Thomas Baker and we need to decide how to responde to these comments
... his seems quite knowledgable and his comments are very interesting
... some comments validate what we hear from Peter
pfps: he is confused about terminology, the interaction between shacl and rdfs and validation
... what I do not understand in his comments is the OWA in the shapes graph
... OWA and CWA have to do with logics but we need to make it clear so people not get confused
hknublau: we closed many of these points no need to reopen them again but there is editorial work to be done
... I am also open to find suitable terms, I am open to renaming
<pfps> I don't think that there is anything in Tom's comments that are not editorial. He is not asking for any changes in SHACL, or at least that is my interpretation of his comments.
arnaud: he is trying to get on board with what we are thiunking but it not very clear
tallted: 1) an editor's draft is not supposed to be consumed from people outside the group, they can look at the next public draft
... 2) reg OWA / CWA they do not apply in what we are doing
arnaud: what we could do is go through thoma's list
pfps: I think people looking at editor's draft are more to be treasured than those looking at the public draft
arnaud: we haven't published in several months and we are due on that
jamsden: let's treat his review like a review from a WG member
<Arnaud> https://lists.w3.org/Archives/Public/public-rdf-shapes/2016May/0000.html
arnaud: the first two points can set the expectations for the reader
pfps: constraint has many meanings, Karen also had problems with the term, we need a better introduction that doesn't include constraint
<jamsden> its not constraints on the graph, its constraints imposed by and application of or use of the graph for a purpose
arnaud: we have two levels of constraints, on a high level of constraining an RDF node with a shape and lower level constraints
jamsden: the word constraint is ok but we do not constrain the graph
kcoyle: we are not constraining we are validating
... we also have components which is confusing
arnaud: we have different use case and we can also constraint web forms
kcoyle: validation is most prominent
marqh: I want to highlight two words, one is "context and the other is "validate", the term validate is not easy to misinterpret
tallted: like other w3c specs we are overloading terms
... validation and constraint can mean the same or different things
... we got to be very clear on the definitions
pfps: every word we use has a meaning associated with it, we should try and use words that are as close as possible to what we want to say
... we either should be very careful when using the term constraint or avoid it. I agree with Karen, validate covers what we do
... why not get away from constraints?
marqh: point 1 in Tom's email is define constraint upfront
... we should also be consistent through out the spec
arnaud: let's go to point 2 "If a shape is described in RDF, say so early on"
pfps: I think there is confusion on the use of rdfs in the shapes graph
arnaud: is this rdf vs rdfs?
pfps: yes, shacl uses only the rdfs terms not the meaning
kcoyle: where does SHACL fit with W3c technology?
arnaud: the problem is with inferencing where we say we do not but sneak some in
<pfps> but what is RDF? is it RDF graphs? is it the old intuitive meaning of RDF graphs? is the the newer formal meaning of RDF graphs? and then what about RDFS?
tallted: classes and subclasses did not originate from RDF, even in biology we have these terms
... someone needs to suggest alternative terminology
marqh: reading section 1.3 can make people worry that something is wrong
... are we trying to be too defensive?
... can we say it in a slightly different way?
tallted: we always need to say there is no reasoning
... you have to do all your reasoning before
<kcoyle> three things: needs an application
marqh: I see a confusion on what is (a shapes graph) and what do I do with it
... can individuals respond to specific comments?
arnaud: I am open, we can certainly to ask for clarifications
pfps: that is not going to work
... otherwise we have N different people saying N different things
... there is a response that I do not completely agree with
arnaud: I have a suggestion, it could be ok for people to ask questions but not to give answers
pfps: email exchange can be very messy and we owe Tom a good response
tallted: we can have a WG delegate do it
arnaud: point 5 suggests we move the extension mechanism to another document
hknublau: this is an opinion of one person
... splitting this makes our job much harder
arnaud: the reason take this seriously is because it validates some comments we had in the WG
tallted: we should take this into consideration but it is only one comment
<hknublau> +1 to a primer if there are volunteers
kcoyle: is this an indication that a primer could be required as a document
arnaud: LDP does something similar
marqh: +1 for a primer document
... people who do not want to implement shacl do not need the whole specification
arnaud: the information is not still stable enough to have a test suite even a primer
kcoyle: I can help with the examples
arnaud: we need to publish another public version of the draft
hknublau: I would like a resolution of datatype vs class before the next draft
... this would be the final syntax changes
... we also have a lot of editorial work
pfps: it would be nice to have an attempt to response to Tom before a public working draft
arnaud: we do not have time to answer all his points one by one before publishing the draft
... I will send a reply to him explaining this
<hknublau> @Dimitris, I'll delete sh:directType
<Arnaud> trackbot, end meeting