Permissions and Obligations Expression Working Group Teleconference

<victor> hello

<victor> I was not here on Thursday. Where can I find the meeting minutes?

<simonstey> in our heads :D

https://www.w3.org/2017/08/24-poe-minutes.html

<simonstey> scribe: simonstey

admin

<renato> https://www.w3.org/2017/08/21-poe-minutes

<michaelS> scribenic: simonstey

renato: approval of last week's minutes

<renato> https://www.w3.org/2017/08/24-poe-minutes.html

renato: minutes of last week's bonus meeting are rather brief

overview of the separate meeting on the 24th

michaelS: we went over the evaluator issues
... ben set the focus on the evaluator document
... we focused on the issue of constraints in a duty
... is a constraint specifying the payamount equal to one that specifies when a duty has to be performed
... i.e. an action might have constraints that apply to actions only
... constraints on action level don't affect whether a rule is active or not

<renato> https://w3c.github.io/poe/model/

renato: based on that, I updated the IM
... I went through and made it clear when constraints where actually "refining" an action, rather constraining the actual rule
... I removed all "in effect" terms

<renato> https://w3c.github.io/poe/model/#constraint

renato: the constraint section was updated accordingly
... also added some subsections and examples
... the constraint section was now moved before the rule section

<renato> https://w3c.github.io/poe/model/#rule

renato: I did add another figure (3) to show how rules can be related to each other

<renato> /w3c.github.io/poe/model/#rule/https://w3c.github.io/poe/model/#rule-active//w3c.github.io/poe/model/#rule

renato: those were the changes that were made according to our recent discussions

michaelS: *grumbles*

<renato> https://w3c.github.io/poe/model/01Constraint.png

michaelS: I noticed that "refinement" looks like 2 words "refi" and "nement"
... we say that every rule can either be active or not active
... all nuances have to be explained clearly

[discussing 2.6.8 in more detail]

michaelS: what deactivates a rule?

renato: in example 25, the duty is a compensate duty

michaelS: so whether a rule is active or not does not take the action of a rule into account?
... so a duty can either be active or not and it can either be fulfilled or not

renato: by default a rule is active (if it doesnt have any constraints attached)

Linda-B: if you have a contract and it has several rules in it
... when are those rules considered to be active?

michaelS: only rules are either active or not, not policies
... policies doesnt have states

renato: policies are groups of rules

Linda-B: ok.. if you have the rules, aren't those rules always "in action/effect"?
... I think we are getting stuck

victor: I'm implementing this and had troubles figuring out whether rules are active
... e.g. once a prohibition was violated, another duty is triggered

<renato> example: https://w3c.github.io/poe/model/#duty-prohib

victor: how can I make it explicit that a rule is active per default?
... I was confused about that.. maybe we should make that explicit in the IM

<victor> "A rule is not active by default if linked to another rule through a failure (consequence or remedy)"

<victor> when i give a url to the failure rules, they do not "fly around"

<victor> my query says "you have two rules"

<michaelS> Why does a remedy de-activate a Prohibition?

<victor> the prohibition is not de-activated, but the remedy is by default inactive

renato: this section def. needs some more editorial work

michaelS: do we finally agree on a view?
... rules are active be default
... if a rule is constrained, all its constraints need to be satisfied for respective rule to become active
... in addition, duties can either be fulfilled or not

Linda-B: if a date is associated and it's before that date, the rule is not active right?

michaelS: yes, as you have an unsatisfied constraint

renato: simonstey, michaelS could you write you point of views down?
... send it to the list

<renato> https://github.com/w3c/poe/projects/1

Road to CR

renato: I hope that's the last "real" issue we have to address
... and we can close off all the other issues as well

ivan: to be "that guy", we have to set a date which for us is the date where all issues are closed and all test cases are up to date and ready
... e.g. 2 meetings from now we "freeze" the spec

<renato> 11th

ivan: if we don't manage to achieve this, we might have to go to the director and tell him we would be able to produce a spec on time
... CR procedure can start at the end of the month
... I've the impression we need at least ~2 months to have implementations ready
... and at that time it's already december

renato: yes +1
... we still haven't gone through the testing regime yet

ivan: we have to do the testing regime asap

renato: well probably need another call for that too
... haven't updated the vocab yet though..

<renato> ok

<renato> ivan: namespace redirects in place for CR

<michaelS> scribenick: renato

ivan: RDFa human-readable form
... maybe by REC
... implementors use Turtle - keep those up to date

michael: Ben's truth tables - are these part of the testing regime

ivan: TT - are they the TT

<michaelS> https://github.com/w3c/poe/projects/1

ivan: existing examples

<michaelS> https://www.w3.org/2016/poe/wiki/Evaluator

ivan: test cases "to be finalised in 2 weeks"....
... namespace and JSON-LD context document

ask Ben to see if we teleconf this Thursday

(30 mins earlier)