Permissions and Obligations Expression Working Group Teleconference

hi!

<renato> Agenda: https://www.w3.org/2016/poe/wiki/Meetings:Telecon20170911

<renato> Chair: renato

i can scribe

<scribe> scribe: victor

<renato> https://www.w3.org/2017/09/04-poe-minutes

Approve last meeting's minutes

<michaelS> scribenick: victor

RESOLUTION: approve last meeting's minutes

Vote on Editors Drafts for Candidate Recommendation

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

renato: the hottest issue is the processing rule section
... ben has been proposing tables (or truth tables) with in-out
... the meaning of "active non active" has been discussed
... ...specially in the context of permissions prohibitions and obligations

michaelS: I have responded to Ben's comments on Friday, no further counter-answer

https://github.com/w3c/poe/issues/254

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

renato: I have some problem with reading the evaluator table, as I dont understand specially the outcome of the evaluation

benws_: the goal is to get the state (active/non-active)

renato: what is the meaning of "an active rule"?

benws_: active permission is I can exercise it
... active prohibition is that I am bound to obey that; obligation the same

michaelS: (about example 22)
...michaelS: I have heard different versions in different calls about this example

benws_: there is a distinction betweens root-level duties and refinements on rules

michaelS: I think have we have the same opinion, I understood it differently in the last call

benws_: constraints in a rule define whether it is active or not, but refinements only "help to".

renato: the truth table needs narrative text

benws_: I think we should limit to the test cases, not give further info to implementors
... Our message to the implementors is: read the im, do the implementation, and pass the tests

renato: do you mean we dont need to speak about rule processing?

benws_: that's it. that section is not needed.

michaelS: We have working long on the Section 2.6.8
... the section has been discussed on github, not finding opposition
... The spec says "an action may include a refinement" but the impact itself on the refinement is not explicitly given
... so the truth tables are going too far

benws_: but without the tables, there is no interoperability

ivan: this is the 4th meeting we have on this. my impression is that there is no consensus in the WG.
... and given that we are in the last minute, we are should take out whatever is conflictive

benws_: ...but if we have disagreement in that section, then we don't have guidedance enough
... I would to have agreement in the tables
... because this is a warranty for interoperability
... we need to define "what is a correct implementation"

ivan: we have a vocabulary supported by the semantics of RDF. This is an RDF vocabulary
... do we want to have additional semantics to what RDF gives me? and if yes, what is?

renato: example?
... "the IM is the semantics"
... we may remove 2.6.a, putting back some sentences taken from other sections in the spec
... what does it mean "a prohibition" is not active?

benws_: that nobody is bound by that prohibition.
... example of temporal restriction: the prohibition is active on Tuesday, but not on Wednesday

renato: so it means "only applies on a Tuesday"?

michaelS: if you violate the prohibition, you have a remedy to fulfill. And then is when the conflict appears.

benws_: if the disagreement is only this small case, we can go further.
... but if we have further disagreement, then we cannot go to CR

victor: I agree with the tables

renato: i have troubles with the terms themselves

ivan: do we have a plan?

CarolineB: active=should be processed, otherwise=can be ignored?

benws_: yes

CarolineB: why dont we choose another word instead of "active/not active"? (just as binary of that, but with other words)

benws_: the previous effort was "in effect / not in effect"

michaelS: Complexity arises from constraints being fullfilled/non-fulfilled, which is a sort of active/non-active state too

benws_: which is the target audience?

CarolineB: it is me

benws_: no! the implementors

CarolineB: but I'll speak to developers

benws_: implementors = developers

LindaB: policy makers must understand the documents and speak to developers

ivan: is the spec correct or no?
... do the truth tables reflect what there is in IM?
... we have 4 months to write a Primer to describe a more human readable text
... are we in the state of going to IM?
... I do not care about the terms (this can be discussed later) but I care

https://www.w3.org/2016/poe/wiki/Validation#Normalization

victor: the 75 policy examples in http://odrlapi.appspot.com/ are very useful

CarolineB: ... is there agreement on the IM ?

benws_: We only fight on example 24

CarolineB: permitted/non-permitted as the equivalent to active non-active in Permissions.

ivan: editorial changes are possible but large ones are not recommended if we go further with the process
... the reading of some specs is incredibly complex, and they can be supplemented by Primers (but we should avoid that if possible)
... benws_ said "example 24 is a small issue" but I see too many issues in the github

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

victor: sI open about 6-8 of them, and all of them "minor"
... but there is no tag for "minor"

renato: we pay attention to the project table instead

<renato> All open issues to be closed

ivan: we cannot go to the director with this amount of open issues

michaelS: what is the impact of constraint and refinement? this is perhaps what is missing in the "constraint property"
... ...this is the section that perhaps can be improved instead.

victor: why don't we have another call Ben/Renato/Michael right now or tomorrow? (and others, i.e., I would like to join as well)

ivan: this change would lead to a second CR

michaelS: it is not a crucial change, it is just "clarifying"

benws_: we make explicit what was implicit

ivan: this is a very fine line

renato: we did take out a couple of sentences from the spec in order to create the "rule processing" section. so it would be simply reverting changes.

https://www.w3.org/2016/poe/wiki/Validation

benws: why don't we meet on Thursday?

<CarolineB> good for me

victor: please review the list of normalization transformations and validations that I have extracted from the specs and implemented. https://www.w3.org/2016/poe/wiki/Validation

michaelS: what is the schedule?
... what about Thursday?

ivan: if it is on Thursday, there must be quorum
... at least a number of people should be there
... besides, I cannot join (even if my presence is unimportant)
... please inform me on the time if you meet

LindaB: 8.30 New York time is fine

ivan: and dont forget closing the issues

<renato> todo: remove section 2.6.8 and add some narrative back into the specific sections

ivan: how much time do we need for the CR?
... the test period, I mean

RESOLUTION: remove section 2.6.8 and add some narrative back into the specific sections

<renato> Next call: Thursday 14 Sept 12:30 GMT

ivan: and on Friday, emails have to be sent

renato: AOB?

michaelS: who will implement the changes in the last resolution?

renato: I will

michaelS: please make sure it is ready by Thursday morning

victor: 22.30 at Renato's, 8.30 Linda's and Victor, 13.30 Ben, 14.30 Michael and Simon

renato: what shall we do with example 22

benws_: oh, no it is correct

michaelS: please read the draft mails

<ivan> trackbot, end telcon