See also: IRC log
<renato> https://www.w3.org/2017/10/09-poe-minutes
renato: no objection, taken as agreed
<renato> https://www.w3.org/2016/poe/track/actions/48
victor: a json-ld document has been created
for each Turtle document
... they are accessibly by the Validation page
<victor> Example
<victor> http://odrlapi.appspot.com/samples/sample005.json
https://www.w3.org/2016/poe/wiki/Validation
<renato> https://www.w3.org/2016/poe/track/actions/49
michaelS: is the Turtle and json-ld numbering in sync
victor: will check
<benws> This test regime has two objectives:
<benws> 1. To ensure that each of the features listed in the Exit Criteria are properly supported by implementations.
<benws> 2. To ensure interoperability between implementations.
<benws> The validation tests help ensure that implementations agree on which policies are valid (and which are not) based on the normative part of the specification. The tests should generate a boolean output (valid/invalid) that matches the values provided in the tests.
<benws> The validation tests do not presuppose the use of SHACL Shapes to test for validity. But shapes covering the tests are provided for those who want to use them.
<benws> The evaluation tests also promote interoperability. Given that two implementations agree as to the state-of-the-world such that they agree on what duties and constraints have been fulfilled and satisfied, then they should agree on what permissions, prohibitions, and obligations an assignee can exercise or be bound by.
benws: has a short outline - and the changes of the Evalution page do note make sense to him any longer.
<renato> https://github.com/w3c/poe/blob/gh-pages/test/test-regime.md
renato: could we agree on Action 49
Discussion about how to use the Implementation Reports
renato: a link to the input sheet should be added to the Test Regime page
benws: don't we presuppose to use SHACL shapes for validation
victor: agreed to that
<renato> https://www.w3.org/2016/poe/wiki/Evaluator
benws: raised issues regarding the updated
Evaluator page
... pointed at example 19
<renato> https://w3c.github.io/poe/model/#prohibition
benws: the truth table does not tell if the
Prohibition is infringed or not
... example 20
... the evalutor does not know that
<simonstey> knowing whether a rule is active or not depends on a blackbox evaluation of optionally attached constraints
<simonstey> the same holds for knowing whether a prohib. is infringed or not
<simonstey> an obligation to not do something is a prohib
michaelS: what should be the result of an evaluator: to check if the results aligns with the rules in the IM definition?
benws: e.g. an obligation not to take an action cannot be fulfilled
<simonstey> a consequence is only active if it was triggered
<simonstey> *might be active
<simonstey> how do you know if a remedy is active?
renato: the Rule column aligns with what we have in the IM
benws: the Information Model is the backgroud for the Evaluator
<victor> I can take a closer look to the Evaluator in the next 10 days
renato: asked benws the bad part of the
truth tables are the Rule columns?
... active is a requirement for setting the state of a Rule
... the evaluator is something we as a group had never a common view on
... it would be the best to integrate the Rule (state) in some way into
the evaluator page
https://w3c.github.io/poe/test/test-regime
<renato> https://w3c.github.io/poe/test/implementors
Agreed: to use the github.io URLs
LindaB: it should be defined soon how to evaluate ODRL
renato: hopes to find an agreement soon - LindaB should tell the group what is currently confusing
<victor> what do you mean more precisely by "Test Dummy"?
<victor> sample input data and output?
<victor> this would be great
renato: remove the testdummy from the implementation page
<renato> https://github.com/w3c/poe/projects/2
michaelS: in the output page, in the input sheet it helps
renato: agreed.
<renato> https://github.com/w3c/poe/issues/278
ivan: the Github issue 278 is still open - how to proceed?
renato: considers it as Editorial - somebody
disagrees?
... nobody raised an objection
ivan: the other issues in the Backlog column
are Editorial - and can be fixed later
... changed #278 to Editorial
ivan: the group has planned to publish some Notes - what are the plans for that?
<victor> I agree with Ben
benws: has plans for a Best Practice and will call in help from Victor and others
victor: for the Formal Semantics Notes he will check with Simon and Sabrina how to proceed
ivan: re IPTC Profile
michaelS: IPTC starts to work on the RightsML Profile, may be published in April 2018
Ivan: when will we have implementations for the implementaiton results
benws: we have implementations but no results for the Evaluator
renato: nothing else to discuss - the meeting is closed.
<ivan> trackbot, end telcon