Provenance Working Group Teleconference -- 02 Aug 2012

<trackbot> Date: 02 August 2012

<pgroth> Scribe: Daniel Garijo

<tlebo> I'll scribe

Admin

<pgroth> http://www.w3.org/2011/prov/meeting/2012-07-26

<tlebo> scribe: tlebo

<pgroth> Minutes of the July 26, 2012 Telecon

<smiles> +1

<TomDN> +1

<Curt> +1

<zednik> +1

proposal: accept minutes

<jun> +1

<hook> +1

<SamCoppens> +1

<stain> +1

<pgroth> accepted: Minutes of the July 26, 2012 Telecon

action items

paul: is still working on his action
... Ivan is joining the WG as W3C POC

ivan: hi!
... sandro will be doing a bit in Aug, but Ivan will be around and back from vacation
... working through all of our documents now.

paul: please sign up for scribes
... aob?

ivan: get unknowns identified

paul: who is P8?

<smiles> or me

<jun> not me

<stain> he knew me last week!

818 is san fran

PROV-Constraints

jcheney: read reviews as they came in. They seem to identify things that we can fix easily.

<pgroth> james?

jcheney: not sig technical changes.

<pgroth> i lost my connection..

go ahead, James

<dgarijo> is there a scribe? I can scribe.

@dgarijo you take scribe

<scribe> scribe: dgarijo

<stain> I am almost finished today

james: waiting for 1 outstanding review from stian

<stain> typing in my evil email..

james: plan is review things early next week in order to be able to have a vote next meeting
... and just worry about minor stuff from there

<tlebo> so, we're voting on this next Thus?

tlebo: are we voting on this next thursday?

pgroth: yes

tlebo: ok, I'll look for responses next week.
... I'll be looking for those

pgroth: in my review there were no blockers

<tlebo> @jcheney they're mini blockers :-)

<stain> I think my review has a tiny blocker .. :-/

<stain> but editorial

pgroth: Issue 458

<ivan> issue-458?

<trackbot> ISSUE-458 -- wasInfluencedBy is not irreflexive -- raised

<trackbot> http://www.w3.org/2011/prov/track/issues/458

pgroth: is influence reflexive or not

<pgroth> http://www.w3.org/2011/prov/track/issues/458

<pgroth> yes

<tlebo> I was the biggest proponent of irreflexive, and I'm no longer protesting to removing it.

<smiles> I agree with removing the constraint

james: does anybody objects to removing that constraint?

<tlebo> (so it's all good)

james: OK, closed

<tlebo> @dgarijo "accepted" is the keyword for a resolution

@tlebo, thanks.

<pgroth> Agent(ex:Luc,[prov;type='prov:Person']) Activity(ex:Luc)

<TomDN> +q

pgroth: the other issue was agent being an activity as well. Should we allow that in the constraints?

<stain> would that imply also Activity(ex:Luc,[prov:type='prov:Person'])

tom: I haven't followed the email about it, but agent can also be an entity.

pgroth: if you assert that Luc is an entity, it can't be an activity.

james: I'm agnostic. It won't break anything to allow people to say that.

<tlebo> from the hip, I think we shouldn't add this.

<stain> agent(ex:workflowEngine, [prov:type='prov:SoftwareAgent']) activity(ex:workflowEngine) would probably make more sense in some cases

simon: I agree with what Tom says. If an agent is a person, then we can infer it's a person.

pgroth: that will imply that all persons are entities.

<TomDN> it could be an inference, and normalization would always produce agent(Luc, [prov:Person]) -> entity(Luc)

<stain> used(food, ex:Luc) wasAssociatedWith(ex:cooking, ex:Luc) wasGeneratedBy(ex:food, ex:cooking)

pgroth: if we want it, then we could add that inference.

tlebo: surprised of the proposal. Is it an issue that went through the list

<stain> yes, should be a proper issue!

pgroth: we can discuss it on the list, but I was wondering if we could just drop it or discuss it a bit here.

<stain> can animals be activities? lifeforms?

<TomDN> machines shouldn't be activities either...

<satya> @TomDN, assuming that machines are "hardware"

<tlebo> why single out Person? Why are Organizations allowed to be Activities?

<stain> a human is in some sense just a system of biological activities working together

jcheney: we should be removing things as much as possible.

<Curt> There was a use case of something like a "working group" which is both an agent and an activity (I disagree with that)

pgroth: even if we don't include something like that I don't know if we can add this after last call.
... Ivan, can we do that.

Ivan: that would be a technical change.

<stain> Curt: but then this all falls back to "are there *any* agents that can be activities?"

Ivan: but constraints is not in last call.

<stain> @Curt

pgroth: so I think james is right. This is going to open up a lot of discussion and given our time line, we should not add it yet.

<Curt> +1 pgroth

<smiles> +1

<tlebo> @pgroth that's a good response: add the explicit type Entity

<stain> I think it would not cause any problems to not say something rather than to say something. It does not affect anything else in the constraints if the prov:Person is an activity or not, and PROV-Constraints is not attempting to model the world, just how processes and things are modelled in a sane way

<TomDN> @pgroth: +1, could we add that as a remark?

pgroth: if anybody wants to assert that persons are not activities, then they can assert it.

+1 to pgroth.

james: I will review questions. I don't have anything more specific

Prox-xml

<Curt> https://docs.google.com/spreadsheet/ccc?key=0Aon9DSj-WtGqdHJocGtPRGp5MTItbEhLX09sN1NxVVE#gid=0

stephan: We are going to try to have xml examples by august 15th.
... hopefully next week.

pgroth: any additional help?

stephan: I'll be able to say next week, when I've taken a stab at it

<Curt> (anyone who wants the doc, just request it through google docs and stephan will approve)

PROV-DC

<pgroth> https://dvcs.w3.org/hg/prov/raw-file/6b795ed2e6c9/dc-note/Overview.html

<ivan> (Curt, I did it)

<tlebo> dgarijo: releaed it last week into hg

<tlebo> ... asking those to reviewed last time to see if they agree to the changes.

<tlebo> .. please raise issues

<tlebo> ... document is not long.

<tlebo> pgroth: reviewers?

<SamCoppens> I will

<smiles> I will review it

<TomDN> I started on a review

thanks!!

<ivan> I will also look at it

<tlebo> ivan: turtle file to put into /ns?

<tlebo> dgarijo: the turtle files is in hg, will send the link.

<tlebo> pgroth: add link to turtle file?

<tlebo> ivan: where to put final version of turtle file?

daniel: I will add the links to the turtel files of each example in the repository. They are there but not linked in the document.

<tlebo> pgroth: the WG also needs to make the full concat for prov.owl with prov-o in it.

<tlebo> @pgroth me too :-)

Handling Comments

pgroth: we discussed last week how to handle public comments
... there was a question about what is a technical comment
... ivan and sandro talked to thomas and have an update.

ivan: the general point is to remind that the last call means that the design is done. Renaming is an editorial change.
... I don't see any problem in this particular case.
... if this change came when we were in CR it would be more problematic.

<tlebo> ivan: changing property name is an editorial change. sees no problem with this. Will need to be more cautious as we go further. name change in CR is impossible b/c it breaks implementations.

<pgroth> http://www.w3.org/2011/prov/wiki/Tracking_Public_Comments

pgroth: the idea behind this wiki page is that we will create a new issue for each comment we receive, to keep good track of them.
... add the issues in the wikipage as well. Is that good enough?

ivan: that will work. Each group decides how to handle this themselves: raise a separate product to deal with comments, to mantain a single table in the wiki, etc.

<tlebo> having a new product seems to make sense, the "dm/o/paq" distinctions are less important now.

<tlebo> Including a link to their email approving our response is a good idea.

pgroth: a new product is fine, but having a last call product for each deliverable would make a busy tracker.

ivan: i don't have a strong opinion, but what you say is right,.

<stain> other trackers have 'versions' for that particular problem

ivan: one more thing: If I look at the wiki page right now I still have to go to each individual issue to see if they are closed or not.
... add for each entry whether the issue is closed in the wiki page.

<stain> I think it should be managable

I agree with Stian.

<jun> whatever works for you, Paul. Fine with either way

not to create a product for each deliverable, it's too much IMO.

pgroth: AOB?

<tlebo> bye!

<pgroth> trackbot, end telcon

- DRAFT -

Provenance Working Group Teleconference

02 Aug 2012

Attendees

Contents