See also: IRC log
<renato> Chair: renato
<renato> Approve minutes: https://www.w3.org/2017/04/10-poe-minutes
<simonstey> scribe: Brian_Ulicny
Approve last week's minutes?
<ivan> scribenick: Brian_Ulicny
<michaelS> scibe: Brian_Ulicny
Approved.
Will we meet next week (May Day)? Decision is to not meet next week.
Resoled: No meeting May 1, 2017.
Resolved.
No feedback received from wide review in the last period.
<renato> https://lists.w3.org/Archives/Public/public-privacy/2017AprJun/0002.html
Horizontal review feedback from privacy group.
Not clear what the issues are. Need to circle back to this group to understand.
Ivan: "this is not about keystroke logging, so seems safe to answer no".
Renato: "OK, I will send that response to privacy group and report back".
Internationalization?
Brian: I was not sure if we were confining ourselves to UTF-8.
Ivan: Yes, we should say UTF-8.
... Where do these issues arise?
Renato: Dublic Core values.
Perhaps.
Brian: URIs or IRIs?
... I only filled out the required field.
fields.
Ivan: the problem is things like
bi-directional text: arabic text with embedded Latin characters. The
problem is you can't encode this in JSON.
... "not sure how the internationalization group handles this these
days"
... "our textual data is metadata. Maybe we can get away with ignoring
this. This may be the only area where there is an issue."
... "They were working on a general approach to this in JSON."
... "The IRI issue is another one. We are essentially an RDF vocabulary.
The RDF spec very clearly refers to IRIs, not URIs. Since this is our
serialization, we should stick with that. We should say that we use IRIs
everywhere."
Renato: "So let's just substitute IRI for
every instance of URI in the spec."
... "So we should be done with internationalization. Brian, please go
over and update based on this discussion and send again."
... "So due date for all feedback is April 30."
... "So that's it for Internationalization."
<renato> https://github.com/w3c/poe/issues?q=is%3Aissue+is%3Aopen+label%3A%22Needs+WG+Decision%22
<renato> https://github.com/w3c/poe/issues/117
Renato: "How to characterize 'Complex Constraints'"
<simonstey> sounds good to me
Renato: "We are going to rename 'Complex
Constraints' as 'Compound Constraints' (cf. 'Atomic Constraints')"
... "Everyone okay with that?"
<michaelS> +1
Silence = agreement.
So 'complex constraints' now 'compound constraints'
<renato> https://github.com/w3c/poe/issues/46
Next issue: in JSON-LD, what does the context file look like?
We've got one context ld file.
<renato> https://github.com/w3c/poe/blob/gh-pages/vocab/ODRL22.jsonld
<simonstey> +q
This file contains all the terms in our ontology. Stuart moved this from his Dropbox.
<simonstey> "action": "odrl:play",
Simon: "should we maintain the odrl
namespace in this example"
... "We need to be consistent"
... "The context file should be more prominent."
... "either we should have odrl:ACTION: odrl:play or
action: play"
<simonstey> "play": "odrl:play", https://github.com/w3c/poe/blob/gh-pages/vocab/ODRL22.jsonld#L96
Ivan: "my experience is that the predicates should be in the context file, but the objects are not guaranteed to be in the namespace, so the example is correct."
"as is".
Simon: "if play is only an object, it shouldn't be in the context file"
Ivan: "yes"
<ivan> http://w3c.github.io/web-annotation/model/wd2/#intended-audience
<ivan> https://www.w3.org/ns/anno.jsonld
<simonstey> (which is the whole point of including a context)
Ivan: "I'm with Michael on this."
<simonstey> https://github.com/w3c/poe/issues/111
Ivan: "We refer to OID in the specs. Let's
try to avoid inconsistency here."
... "let's clean up these minor inconsistencies."
"let's put the cleaned up json file"
"online."
"let's look at a specific example."
<renato> https://w3c.github.io/poe/model/#policy
<renato> Example 1
<simonstey> s/"@context": { "odrl": "http://www.w3.org/ns/odrl/2/"/"@context": "http://www.w3.org/ns/odrl.jsonld",/
<ivan> "@context" : "https://www.w3.org/ns/????.jsonld",
<simonstey> s/"@type": "odrl:Set", /"type": "Set",/
<ivan> "type" : "Set"
"This is how the context statement should be"
<ivan> "action": "read",
"ID or OID also"
"Should show to Greg Kellog ? to be sure."
"Need to be consistent with ID or UID"
Ivan: "is it okay to put this in /ns?"
<renato> https://www.w3.org/ns/odrl.jsonld
Renato: "Yes, and we will change the URIs"
... "Need confirmation from Stuart"
Ivan: "Yes, and this can change at any time."
<renato> https://github.com/w3c/poe/issues/139
REnato: "two proposals to deprecate two section"
<simonstey> +1
"how to deal with undefined actions. This section doesn't add much value."
<simonstey> also relevant https://github.com/w3c/poe/issues/105
"what should you do when you come across a previously unseen action?"
<ivan> ODRL jsonld context is in place: https://www.w3.org/ns/odrl.jsonld
<renato> https://w3c.github.io/poe/model/#undefined
Ivan: "3.1.4 - undefined action is different than unsupported action"
"what is it to be _defined_ action?"
REnato: "WE need to decide how strongly we want to embrace this, or remove it?"
<simonstey> +q
Simon: "in favor of removing this section. In favor of relating new terms to previous ones."
"let's define dependencies among actions."
Michael: "wouldn't this be a job for the definition of the action, to specify its relations?"
Simon: "this would be similar to SKOS
vocabulary broader-than, etc."
... "adding actions is not part of a profile"
<renato> https://github.com/w3c/poe/issues/138
<renato> https://w3c.github.io/poe/model/#conflict
<simonstey> +q
Renato: "For next time, conflict policies"
Simon: "this needs a lot of discussion. Totally against removing it."
"print more than 10 times and print less than 5 times. ..."
"needs a lot of discussion."
AOB?
meeting adjourned. Next meeting in 2 weeks.
<ivan> trackbot, end telcon