See also: IRC log
<phila> scribe: Brian
<phila> scribeNick: Brian_Ulicny
RESOLUTION: Accept last week's minutes https://www.w3.org/2016/10/10-poe-minutes
Simon has gone through all the requirements based on the TPAC.
benws: "the list of requirements on the sheet were the ones that had been accepted"
<simonstey> https://www.w3.org/2016/poe/wiki/Requirements
Simon asks the group before the next call to go through the wiki requirements and check whether Simon has correctly captured the status of the requirements.
<victor> ok, so we are supposed to read this again and provide feedback.
Simon directs attention especially to the blue colored requirements.
So group is enjoined to please look at the use case doc and requirements.
<simonstey> +q
Michael mentions that the contributor of the use case should make sure that the use cases and requirements line up.
Michael "especially the ones marked agreed without discussion"
benws: "how could they be agreed without discussion?"
michael: "they were assumed to be agreed based on call to discuss"
Phil: "am I supposed to be looking at wiki or editor's draft?"
<michaelS> q
Simon: "look at wiki"
benws: "agreed that wiki is canonical version"
<benws> "for now ...2
phil: do we need to assign action items? Or agree that after a certain date, everything will be transferred from wiki to editor's draft.
"we need to concentrate the mind"
simon: "there's only one requirement left that has 'agreed without discussion'"
has to do with notes and formal semantics
michael: "on lowest part of page, we have a 'needs consideration'"
"three have undefined status -- only proposed"
"so are they still under consideration"
Simon: "no, I forgot to change status. Discussed on 23d. Will change"
benws: "let's move on to tech discussion - constraints on constraints"
"three possibilities"
"let's take use case of embargo"
"one suggestion: create an emargo"
"another solution (ivan): create two kinds of has constraint property -- one like we already have, another a qualified containt"
"third option: chain constraints together"
"any other options?"
victor: "sent an example earlier this am as Sparql query"
<victor> http://cosasbuenas.es/static/ComplexrestrictionswithSPARQL.pdf
victor posts suggestion
victor: "requires sparql endpoint"
<simonstey> which is actually SHACL
"using SPIN"
"example is self-explanatory"
benws: "isn't this an example of the first use case - creation of a complex object"
victor: "Yes that's right"
michael: "i could only join meeting by audio in lisbon"
"can't understand model behind these examples"
"would be great help for everyone on the call"
benws: "I will publish some example"
"existing constraint model allows contraint only on one predicate or property"
"let's say there's an embargo on extract from football match"
"need to know end of event time and offset interval"
"our existing constraint model can only handle one parameter"
benws: "the first model is where object of predicate is a complex item -- e.g. an embargo object"
"second option allows two types of constraint: one standard, the second is qualified constraint qualified by external object"
"inject the additional constraints into the external object"
"third option: multiple constraints chained together"
simon: "have a lot to say"
"before we introduce fancy constraints, we need to understand what we want ODRL to be/allow"
"do we want to do constraint checking?"
"do we want to verify that constraints hold?"
simon: are we just concerned with expressiveness?
<simonstey> http://w3c.github.io/data-shapes/shacl/
"SHACL working group output could be used to express constraints"
"if we go with chaining, which looks attractive,...there are problems"
"looping, for example"
"i really want to discourage going down chaining route"
phil: "whichever option is chosen, need to build it"
"someone should try to implement this stuff -- W3C suggests that you need to do this stuff"
benws: "I am writing code around this"
<Zakim> phila, you wanted to say that any one of those three objections on their own out this solution
benws: "can simon provide some concrete examples of dangers of chaining?"
simon: "yes"
phil: "problems listed in the PDF show why this solution is not the right one"
sm: "plea for non-complex constraints"
"video editors have these constraints all the time"
"often an embargo is turned from relative time to concrete time: embargo will definitely be over by x o'clock"
<simonstey> +q
benws: "embargoes are not the only complex constraint"
"payments are another example"
simon: "regarding renato's chaining example, not sure if example is complex enough"
<simonstey> odrl:constraint [ a odrl:Constraint ; odrl:operator odrl:eq ; odrl:event <http://epl.com/vocab/afterMatch> ; odrl:constraint <http://example.com/30minEmbargo>; ]. <http://example.com/30minEmbargo> a odrl:Constraint ; odrl:operator odrl:eq ; odrl:dateTime "P30M"^^xsd:duration .
"permission is granted during an event"
"point is that I've encountered a situation where there is no property that can be used"
"the property is not adequate to address the use case"
benws: "let's take this discussion to mailing list and reconvene next week"
"will provide concrete examples"
<simonstey> +q
<simonstey> https://github.com/w3c/poe/issues/45
simon: "raised an issue on github wrt relations"
"does it make sense to have extended relations for permissions?"
<simonstey> https://w3c.github.io/poe/model/#extended-relations
"spec says that if you and permissions then you have to do all of them. this is wrong"
"you have to do all ANDed duties, not permission"
"permissions"
"what does it mean to tie permissions with an OR?"
simon: "either you have to satisfy this constraint or that constraint"
"specification is wrong. Uses 'may not', which is not defined."
serena: "I agree with simon. Connectors for duties and constraints, not relations"
<phila> PROPOSED: That the extended relations (AND, OR, XOR) will only apply to Duties and Constraints, not Permissions and Prohibitions
<simonstey> +1
<benws> +1
<Serena> +1
not sure
<James> +1
<phila> victor: +1
RESOLUTION: That the extended relations (AND, OR, XOR) will only apply to Duties and Constraints, not Permissions and Prohibitions
<smyles> +1
<michaelS> +1
benws: "will quickly run through open actions"
<benws> https://www.w3.org/2016/poe/track/actions/open
<phila> action-25?
<trackbot> action-25 -- Víctor Rodríguez-Doncel to Provide an example of how good relations vocab support 'unit-of-count' -- due 2016-09-29 -- OPEN
<trackbot> https://www.w3.org/2016/poe/track/actions/25
"victor, you were going to provide an example of unit-of-count"
<phila> action-30?
<trackbot> action-30 -- Stuart Myles to Can we only have a json-ld serialisation? will it impact righstml? -- due 2016-09-30 -- OPEN
<trackbot> https://www.w3.org/2016/poe/track/actions/30
"simon, yours will always be open"
"issues raised"
phila: "do we need to follow issues in github as well?"
"we just need to make sure that we track ALL the issues"
simon: "is there integration with IRC and the github tracker?"
phila: "No"
<phila> GitHub Tracker
15 issues on github tracker currently
phila: "Most raised by renator"
*renato
<simonstey> issue-1
<trackbot> issue-1 -- The number of times we need to refer to the target -- closed
<trackbot> https://www.w3.org/2016/poe/track/issues/1
simon: "easier to look up issues in IRC via trackbot"
<victor> I can solve mine inmediatly Action 25. the example that is requested can be seen in slide 18 of http://tutorials.oeg-upm.net/rightslinkeddata/session4b.pdf
benws: "can bring this up next week."
victor: "next f2f: when, where, who (ben is necessary)"
<phila> phila: I think Renato might well be there
benws: "suggest a date, we'll take it from there"