Permissions and Obligations Expression Working Group Teleconference
03 April 2017
<renato> https://www.w3.org/2017/03/27-poe-minutes
<phila> [Nothing heard]
Any comments on the last minutes, otherwise I take silence as approval
Accepted
Resolved: Last week's minutes accepted
Update on the wide and horizontal reviews
<renato> https://www.w3.org/2016/poe/wiki/Deliverables
Renato: Just waiting for feedback. Ben did you speak to Brian about the internationalisation
Michael: Would it help to indicate who has replied. I have done this.
<renato> https://github.com/w3c/poe/issues/118
Renato: Probably a good idea. I have done this the other way around from the issue. Good idea to populate both lists
renato: Back tht einternationalisation
benws_: I haven't had a chance yet. I will followup.
renato: Nothing else to report
renato: After we get the feedback we need to setup things in order to get to CR e.g. leader for testing
phila: We haven't talked about it in a while
… we need independent implementation that work the same way - one way is to setup a test suite - including inputs and outputs
… This can be another note or maybe even a github repo
… A set of machine readable policies and demonstrate that there is more that one implementation - given the same input we get the same output
Renato: How would that look for us given we don't do any processing
benws_: We are using a validator for the terms that we use... not the whole vocabulary
phila: Each individual piece of software does not need to cover everything
benws_: Do we need to make this public
phila: No it doesn't need to be public, you just need to give us the information
smyles: We had an ODRL engine with 3 things for testing - 1) the policy, 2) the result (yes, no or maybe) and 3) the context (who is the assignee, what country or city are they from, what action are they trying to perform)
… This was enough for us
benws_: It sounds like you were testing a processing model. However we need to syntactically validate policies.
<phila> "This will require the ODRL Policy to be interpreted with the additional information identified by the URI."
benws_: is that correct Phil
phila: Just did a search for interpret - points to a concrete example - how one choses to use this information is out of scope for the working group
… this is perfectly ok, however the uses need to be consistent and we must show they are consistent with one another
<renato> The processing model for Constraint Relations includes:
benws_: Sounds like a processing model
renato: Points to the 9 steps listed in the processing model
<renato> https://w3c.github.io/poe/model/#constraint-relations
renato: Put up a test e.g. to say expressions need to be atomic
… There are a few points were we can pull out a processing model and turn them into a text case
benws_: We can check is this a valid piece of ODRL and does it allow me to take this action
… the 2nd is more enforcement
… I guess we need to avoid this
phila: Apart from being valid RDF, what is valid ORDL
benws_: Enforcement of the information model
phila: Yes they are different.
<phila> ShEx or SHACL
phila: Vaidation could be done in various ways 1) Shacl and 2) ShEx
The group proposes to the director what the exit criteria should be
Phil: You could suggest multiple instances (at least twice) of all of the terms
… For core terms they should of course be used much more
… Plus examples of real live ODRL policies
… It would also be good to indicate the type of uses we've heard about that would be good aswell (e..g what Ben and Stewart are doing)
… Details of real world valid use
benws_: What needs to be in the tests
phila: All the terms and all the constraints
… build some kind of software and test against it
renato: ... build the tests including invalid tests
phila: yes
renato: does the test suite need to be complete before the F2F
… How much needs to be complete by May
phila: It doesn't need to be complete, it should be started
… with the experience in this group it should not take too long
… plus you might find that you need to add to it it later when you are in CR
smyles: I would like to remind the group that ODRL is not just RDF
… We will also need to do XML validation
<phila> +1 to smyles
smyles: It would be useful to evaluate engines that process ODRL and when 2 different engines process ODRL they come to the same answer
… this would be ambitious by May but ultimately very helpful
renato: It depends on how complete it needs to be
… however this would give the director confidence
… we need to ask the group if someone would like to lead this piece of work
… so that we have some test cases for the F2F meeting in May
smyles: Is the challenge organising the work or doing the testing?
renato: This is really about managing the process (getting the test cases ready to go)
phila: I could have a go... I'm not sure if I can do the whole thing but I could certainly write some examples
… Sabrina are building a parser as part of SPECIAL
… We are modelling at the moment, so if depends on how that goes
renato: Thanks Phil for kicking that off
… any other questions on the testing regime?
Nope
Any update Simon on the formal semantics
<simonstey> will have a call in the next weeks
<phila> Sabrina: We have exchanged some e-mails about formal semantics. We'll kick off next week when Ivan is back from WWW in Perth
renato: Ben any update on the best practices
<renato> https://www.w3.org/2016/poe/track/actions/open
renato: Ben and Phil and update on your actions
phila: No update on the accessibility usecase
smyles: Question on versioning
<renato> https://www.w3.org/2016/poe/wiki/Meetings:London2017
benws_: I can confirm the address is correct
renato: I have prepared a draft agenda... wide review, horizontal review etc...
benws_: Renato will you attend in person?
renato: I'm not sure at this point
michaelS: There are currently 18 issues, however it is not clear how these issues can be resolved
<michaelS> 18 open issues
renato: We should talk about some of these next week - especially the ones that say need working group decision
… we could maybe tackle 2 per call
<renato> https://github.com/w3c/poe/issues/22
renato: example is number 22
… proposal from Simon to remove the inherit relation
… this is probably something that the editors can not decide as it is really a group decision
phila: All of these need to be closed before seeking CR transition
… one relates to the usecase document. Do we plan to do more work on the usecase document? If so we can do this and republish the usecase document. If not we should close it and say why.
renato: We hope to close as many as possible before the F2F
<simonstey> I'll address the ucr issue until next week
renato: it would be good to at least discuss them before the F2F
<simonstey> I encourage everyone to read through the spec and raise issues themselves
renato: I would like to propose a couple of github issues per week for discussion on the weekly call
… will start with the ones that say need working group decision
<renato> IPTC
Phil: I have a talk at the IPTC, will anyone else be there?
… I will endeavour to prepare the presentation in advance and get your feedback
smyles: I have been briefing the IPTC group for sometime on ODRL and now POE
michaelS: I think it would help to clarify what the W3C review requirements are, the IPTC people will not be familiar with these
<simonstey> https://www.w3.org/2008/04/scribe.html