See also: IRC log
<BartvanLeeuwen> Zakim: ??P6 is me
<Dimitris> zakim jimkont is me
<scribe> scribe: hsolbrig
Arnaud: new Scribe.perl not in cvs yet. EricP will submit
... generated minutes now have a summary of resolutions
... with ids on resolutions so they can be linked to
PROPOSED: Approve minutes of the 11 December Telecon: http://www.w3.org/2014/12/11-shapes-minutes.html
<pfps> Yay that we can now point to resolutions in minutes.
<pfps> The minutes from last week look OK.
Minutes approved
RESOLUTION: Approve minutes of the 11 December Telecon: http://www.w3.org/2014/12/11-shapes-minutes.html
Next meeting on January 8, 2015
<ericP> scribenick: hsolbrig
<SimonSteyskal> you have to prepend the name of the person who's is speaking
Action 5 is completed and wiki page has been created for requirements
Arnaud: Stories page is frozen. New stories will only be insterted if approved by the group.
PROPOSED: "Freeze" and approve the user stories we've collected as of 18 December 2014
pfps: I think there is a lot that isn't appropriate to approve in that document.
... Happy to say that these are our stories, but it is still possible to challenge story and perhaps drop.
... Comments in documents have no responses.
Arnaud: we should turn comments into issues and discuss
PROPOSED: "Freeze" the user stories we've collected as of 18 December 2014
Arnaud: Frozen stories are considered in proposed stage. People should raise separate issues for every story they have an issue with.
<pfps> I'm happy raising issues for stories that are not in what I consider an acceptable state.
<SimonSteyskal> +1
+1
<pfps> +1
<magyarblip> +1
<kcoyle> +1
<ericP> +1
<BartvanLeeuwen> +1
<Dimitris> +1
RESOLUTION: "Freeze" the user stories we've collected as of 18 December 2014
<pfps> I'm fine with "approved by default"
Arnaud: Default will be to approve user story unless an issue is raised.
Arnaud: pfps and hknublau created requirements wiki page (https://www.w3.org/2014/data-shapes/wiki/Requirements)
... Disagreement on how we get from proposed to approved. Is intermediate step required?
pfps: Propose "contract" - if requirement is under consideration, every working group member has to understand it. Needs to be an intial step -- "I think this could be a requirement..."
ericP: One person has to understand it before it moves to under consideration...
pfps: DC had 200+ requirements to be evaluated. Need some way to triage
Arnaud: How do we determine getting things moved to under consideration?
pfps: Need to be sure that every requirement gets serious examination and we don't have 200 all at once.
... PROPOSED: Every member reviews requirement and it needs least one second before it goes to a vote.
<SimonSteyskal> http://www.w3.org/2000/09/dbwg/details?group=73865&public=1
<SimonSteyskal> ;)
Arnaud: Requirements have someones name associated with them. People endorse requirements by adding their name to them threshold of 2-3 names (2 too low?)
<magyarblip> how about ensuring the seconder is semantically distinct from the proposer?
<pfps> I'm fine with endorsements coming via tags on the wiki page. Maybe then there should be both positive and negative votes?
ericP: Issue with requirements changing while being clarified.
... Voting includes refinement?
pfps: Things could get dropped back to proposed stage if there is confusion or voters were mistaken about what they voted for.
hknublau: I'm happy with any process as long as there is reasonable timeline. What is role of stories and requriements in big picture? At what stage do we address proposed solutions? When do we shift to actual work?
Arnaud: Wish we could go faster. Process is evaluate stories and requirements then look at proposed solutions. Not sure how to go faster without folks feeling that things are not fair.
hknoblau: Requirements need to developed along with solution. We could create thousands of requirements, but solution may explode. Solution must be concise and sufficient for the job.
hnkoblau: If we delete requirements too early we may throw away good things
Arnaud: Suggest we start looking at the requirements and read through them, people can add their initials to them. Third person moves section to under consideration. In a week or two we should be able to start talking about proposed solutions
hnkoblau: suggest "+" or "-" that allows people to express their concern
<pfps> I'm not against keeping the requirements in place, but there should be some better way of tagging them. Setting up classes would be fine, as long as it was easy to see the status of the requirements.
arnaud: We can discuss the write method to tag things, but from a process perspective I feel this is reasonable
PROPOSED: People should add their name (initials) to requirements they think ought to be considered, if there are three persons endorsing a requirement it is then "under consideration"
+1
<SimonSteyskal> +1
<hknublau> +1
<pfps> +1
<Dimitris> +1
<ArthurRyman> +1
<kcoyle> +1
<BartvanLeeuwen> +1
RESOLUTION: People should add their name (initials) to requirements they think ought to be considered, if there are at least three persons endorsing a requirement it is then "under consideration"
Arnaud: We cannot allow this to go on for too long. Tag soon!
<ericP> +1
pfps: Feel free to comment without voting + or -
pfps: Two issues. First is use of the word "Resource". Dick Cheney is a "Resource", working group is not trying to find Dick Cheney's shape.
ericP: Even if we misuse "resource" to refer to graph nodes, we still need to make this publishable.
ArthurRyman: May want to describe the shape of an information resource that describes Dick Cheney.
pfps: What is the "shape" of a web page?
ArthurRyman: We are talking about rdf information resources not web pages. An RDF graph is a representation of an "information resource", you can GET through http. We are concerned about RDF information resources
pfps: The RDF representation of an information resource is a graph that *describes* a resource.
ArthurRyman: you never get an information resource you get an representation of it.
ArthurRyam: You may have a URL that is a picture of Dick Cheney. You can negotiate for format (.png , gif, etc).
ArthurRyman: Shape is a term that is used to describe expected content and constraints of RDF Graph
pfps: Two different communities in this working group
ArthurRyman: Personally I'm interested in linked data aspect, others in community are more interested in pure RDF. Need to be relevant to linked data world.
Arnaud: Peter started a glossary -- maybe we should hash this out offline and add definitions to the glossary?
pfps: If anyone is going to refer to "Resource" they need to deliniate what it is or isn't. Even "web resource" isn't sufficient
<Zakim> ericP, you wanted to make a plea for a term in a graph which has properties conforming to some "shape"
ericP: Objective is to come up with some term that describes a "term in a graph" a subject or object of a set of triples and a pattern around it. Do we agree?
pfps: We have "Node", which includes literals.
ArthurRyman: We can have an identifier for Dick Cheney, and that can be a node in a graph. What is the objection?
pfps: Is the node unrepentent?
... Need to differentiate what software can work with (graph) and what it can't (Dick)
+q
<ericP> hsolbrig: i agree with pfps's point. ran into issues when working with ericP on the [ShEx] draft about separating nodes from their dentation.
<ericP> ... a node could be about dick cheynei, but for shapes, it's just a node.
ArthurRyman: We're defining shapes for nodes in a graph. Shape describes graph, not just a node...
Is he the good twin?
<ericP> i now believe "rooted graph" is the better understood term of art. i got "pointed graph" working with someone who translated it from french.
ArthurRyman: You could say that a node on a graph that contains 12 arcs. Real world examples of disconnected graphs exist.
... Json doesn't require things in JSON/ JSON LD to be completely connected.
<pfps> For these situations to be considered, I would very much like to have some stories and use cases for them.
Arnaud: I think Peter objects to "resource" because we are inheriting from RDF. Maybe we need to invent a placeholder term to facilitate the discussion?
pfps: Don't know what the target is supposed to be. What I think we're talking about already has a perfectly good RDF name.
<Zakim> ericP, you wanted to nominate pfps anyways
<ArthurRyman> SEE s35 https://www.w3.org/2014/data-shapes/wiki/User_Stories#S35:_Describe_disconnected_graphs
ericP: I suspect that Arthur wants a term for a set or island of data.
ArthurRyman: I don't see conflict, we just need to describe what we are talking about. Suggest 1-1 call
pfps: What else is there besides "node" and "rdf/graph"
Arnaud: We need an answer for the next call. Peter and Arthur need to duke it out and have an answer to the thing we're talking about in issue 5.
<ericP> i suspect that disconnected graphs are implicitly connected, and that should be made explicit before validation.
<BartvanLeeuwen> bye
<pfps> Thanks to Arnaud for fixup up the minutes script.
<Arnaud> trackbot, end meeting