Provenance Working Group Teleconference -- 23 Feb 2012

<trackbot> Date: 23 February 2012

<Luc> Scribe: stain

<Luc> hi Stian, thanks for volunteering!

<Luc> @macted, any feedback on prov-dm proposed restructuring. Can you share some feedback?

<pgroth> hi sandro are you on today?

<Luc> hi stian, it's all set up, are you ready?

yes

<Luc> great, thanks for volunteering

not sure if zakim recognized me, but that's not important

admin

<Luc> paul, should we get f2f2 minutes approved today?

<Paolo> for SIP users: can you connect to zakim@voip.w3.org?? I can't

Luc: Call now starting.
... Review PROV-DM and PROV-O
... release of documents.. if time, we'll look at proposal for binary relations for 5th working draft (of DM?)

<Luc> PROPOSED: to approve the minutes of Feb 16 2012 Telecon: http://www.w3.org/2011/prov/meeting/2012-02-16

<tlebo> +1

<khalidbelhajjame> +1

<ericstephan> +1

<Curt> +1

<GK> ABSTAIN - NOT SEEN THEM YET

<stephenc> +1

<pgroth> +q to comment on f2f minutes

<Luc> ACCEPTED: the minutes of Feb 16 2012 Telecon: http://www.w3.org/2011/prov/meeting/2012-02-16

Paul: Just finished the minutes - but we can't do approval now as people have not read it yet
... the minutes of the F2F2
... apologies for delay

Review of actions

Luc: Action-55 was reopened to complete OWL file - this seems now done and can be closed. We'll review it.
... Action on Paul to propose proposal,

Paul: Talked about it last week, and to talk about it in two weeks time (ie. next week?)

is that 2 weeks from today or last week?

<Zakim> pgroth, you wanted to comment on f2f minutes and to comment on f2f minutes

Luc: ACTION-61 to update prov-sem

James: Travelling next week, so will have it done before then, not yet done

<GK> @paul I should be in a position to be a little responsive on PAQ issues next week

^^.. action on Paul was ACTION-57

Luc: ACTION-63 Structure of HTML file for PROV-O document - postponed

PROV-DM Simplification: Reviewer feedback

Luc: Feedback on PROV-DM simplification. Last week we released 3 separate documents, one called PROV-DM, one PROV-DM constraints, and one PROV-ASN
... We lined up reviewers and invited for review of docs
... to identify/decide a number of issues that are in the agenda

<GK> In agenda: http://lists.w3.org/Archives/Public/public-prov-wg/2012Feb/0415.html - but I reviewed the wrong document; I've just posted a brief update

http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.02.23#PROV-DM_Simplification:_Reviewer_feedback

Luc: Try to reach consensus - if possible - links to emails sent by reviewers
... perhaps a quick summary from each of them?
... about if restructuring of docs are addressing points
... Tim first

Tim: Feel that new draft has dramatically adressed the concerns.
... Sent email this morning with detailed comments

Luc: Missed link to that email

<pgroth> http://lists.w3.org/Archives/Public/public-prov-wg/2012Feb/0416.html

<pgroth> i'll edit the agenda

Eric?

Eric: Document was over-all, great job of meeting simplification objective

Daniel?

Daniel: Have not finished whole document, made it to the middle.. made some notes that I was planning to send
... Try to take my W3C hat off, and try to identify what is confusing to me
... Will send small details in separate emails

Luc: What about meeting simplification objectives from F2F?

Daniel: Think that it more or less has accomplished this, but not gone through the whole doc. Much clearer now.

Luc: MacTed? Might not be on call yet

<pgroth> MacTed?

Curt: First part easier to read, many things still confuse me. Second and Third, mechanics work well.

(??)

Sam: Find the overall structure very clear, nice separation of concerns

conserns

Sam: All 3 well written. Sent list of some remarks. (to whome?)
... has also reviewed part 2 and 3, which I'll send

<tlebo> (back onto IRC, @luc, my email with comments is http://www.w3.org/mid/995BD58C-DB94-4052-BE85-BE9A271695C0@rpi.edu )

<pgroth> sam I don't see your email

Sam: Can recognize this person to become editor of draft

<SamCoppens> Excuse me, I have sent it to Luc

Jun: First time I read this document - did not read previous version, and so have no comparison

SamCoppens: sorry :

Jun: To summarise, don't think the simplify document.. (?)
... Not ready for editorial draft at the moment
... 1) Lack of context and explanation
... Reading it for the first time it was difficult to follow

<SamCoppens> My remark was for Paul

<tlebo> glad we're getting @jun's fresh eyes :-)

Jun: Second paul I want to say is, I did not make a clean/clear explanation about.. provenance.
... I'm just referring to minutes of F2F meeting
... not exchanged in current draft (?)
... Does not help me explain how this reach the new goal.
... Luc might tell me how this structure, part1/part2/part3, how it is reflected in part 1

<pgroth> jun which document did you read?

(??)

(I'm very confused)

Luc: Many things not consistent -f or instance figure not consistent with section with overview

^^ Jun:

Jun: Mixed terminology, elements/edges/properties/classes
... I don't mind which terminology we use, as long as it's used precisely, but that is not the case in this document
... There's lots of references to other sections not existing anymore, terminology that might become obsolete.. too
... too early to raise comments on those now?
... Perhaps focus on something different?
... Interested in Luc's feedback

<Luc> http://www.w3.org/TR/2012/WD-prov-dm-20120202/

Luc: Have responded to your email. We'll ask the reviewers what they have addressed.
... ^^ is the second working draft
... what we are standing is wether the document as it stands can be used as an editors draft

(is that a different document?)

<pgroth> point of clarification

Luc: If you believe that we should not do this, then what are the blocking issues form your point

Jun: Not quite covered in my email - how this new structure corresponds to the scruffy and precise notation

Luc: I think these are terms we've used informally, not used specifically

<Curt> scruffy = you forgot to read part II

<GK> (I agree that "scruffy" and "precise" are informal)

Luc: We have defined a vocabulary, those using the vocabulary will make scruffy provenance

<pgroth> +1 curt

Luc: If you follow the constraints of part 2, then it is a more refined provenance, more precise about what it is asserting.

Jun: So you are saying that this new working draft is related to an even longer document..?

<Paolo> @Jun: yes!

Luc: Yes, all those 3 documents were 1 big document
... We've tried to also simplify the presentation

Paolo?

Paolo: Trying to locate an email I sent to Jun..

<Luc> @Curt, I like this!

Paolo: Main point is that according to the process/goals we put in place at F2F
... simplify what was there
... question is, what that achived to an extent that we can discard the previous version
... and use this as a new baseline
... that is the question.
... So Jun, I would ask you to look at the current baseline with that perspective - which is different than coming from blank
... we're aware that that's what you promised.. so question is, is this a sufficiently good baseline
... but then you need to know what the old massive document was
... in my email, this scruffy vs proper is a placeholder to say is there something we can isolate as essential (part 1) and the rest in part 2.
... this split should give a simplification - not labelling everything as scruffy or proper
... just a way to encode a progression from simplest possible to be useful, to more sophisticated use
... That is the email I think I sent 30 minutes ago

Jun: I think you managed to convince me, I must apologize. Where we started is this massive long document.
... so this is an encouraging first step. And I hope my comments can be used for consideration further in the editorial process
... So YES, it could be a baseline for further work

Paolo: Some things pointed out not taken into account - like what is this about. That is coming.

pgroth: about process..
... we've had pretty sophisticated reviews, need to figure out how to distill these to editorial issues, and 'real' issues on concepts

GK?

<GK> http://lists.w3.org/Archives/Public/public-prov-wg/2012Feb/0448.html

GK: spent all morning reviewing the wrong document
... posted a brief update ^^
... first comment: New document is definetly moving in right direction
... some comments from my review this morning still apply
... but many have been addressed, so I think this is something we can build on
... rest of the issues are technical issues
... which we'll discuss as we get on with it

Luc: Sorry you spent so much time reviewing WD3 - the wrong document
... for the working draft it would be good to get a number of resolutions approved
... have anyone else reviewed the documents and want to provide feedback?

(Stian: I've had a quick look at part 1, which looks good, but no review)

Luc: Want a clear statement from working group that we want the document split into 3
... we need to do this to get a transition request to get the new documents approved

<GK> Just to clarify: I think there are both editorial and technical issues to address in DM

Luc: need to work with sandro and ivan to make a strong case for W3C
... to have that resolution agreed..

pgroth: is that really the case?

<Curt> Is this 3 documents or 3 parts of 1 document?

pgroth: Sandro?

sandro: not a strong case.. if the WG resolves that it's the right thing to do, we can make it happen

<Luc> PROPOSED: The Working Group supports the restructuring of the PROV-DM deliverable into three separate documents, currently named PROV-DM, PROV-DM-CONSTRAINTS, and PROV-ASN to facilitate its presentation.

sandro: question is what happens with the older one.. should these have the same URLs?

<jun> @curt and gk, that's what confused me:) and now i understood their relationship

Luc: Propose to keep same name for PROV-DM

sandro: yes, that solves that issue

Luc: propose two new names.

sandro: just consider them as new working drafts

Luc: but procedurally we need to make sure it's the same deliverable, for recommendations, etc

<Curt> 3 URLs = 3 html documents

sandro: yes, same deliverable in 3 documents

Luc: if we are happy with this proposal, can you express your support?

Paul?

(?)

GK: Do we need to split it into 3 documents?
... Division of material in part 1, part 2 in particular, (part 3 is useful), do we then need 3 separate documents? OR structure it within a single document?

<pgroth> +q

Luc: my recommendation as editor is 3 documents

<pgroth> +1 to 3 documents

Luc: which gives the entry points to DM much lighter
... many are not interested in constraints, just want a description
... a long document is daunthing
... external feedback from Tom Baker and IVan both suggest splitting deliverable in separate documents

<Zakim> GK, you wanted to ask: can we have part1, part2, part3 in the same document?

<Curt> +1 make 3 separate documents, include introduction/scope in each describing their relationship clearly

<jcheney> afk, supportive of splitting (for now at least)

<Paolo> +1 for spliitng

pgroth: also think we should have 3 docs

hang on

<pgroth> +q

stain: could it not just be 3 html pages on one document (same base URI)?
... some recommendations do that

<GK> @paul +1 (easiest way in w3c process; editor's discretion)

stain: if it is to be 3 separate documents, then they should be valuable on its own, say referring to PROV-DM-CONSTRAINTS alone. Don't have a view if that 's the case or not

Luc: we can come back to working group if needed

<pgroth> good with me

Luc: Express your support

<dgarijo> +1

<smiles> +1

<GK> +1

<Paolo> +1

(or not)

<ericstephan> +1

<Curt> +1

<SamCoppens> +1+1

<sandro> +1

<satya> +1

<khalidbelhajjame> +1

<zednik_> +1

<jun> +1

<Luc> ACCEPTED: The Working Group supports the restructuring of the PROV-DM deliverable into three separate documents, currently named PROV-DM, PROV-DM-CONSTRAINTS, and PROV-ASN to facilitate its presentation.

Luc: Second point is to agree or not if the document as it stands can become editorial draft

<Luc> PROPOSED: the three current documents PROV-DM, PROV-DM-CONSTRAINTS, and PROV-ASN should become the new Editor's draft.

Luc: that does not mean we have to release them as editors draft.. but they are the current editors draft according to w3c terminology

<Paolo> +1

<SamCoppens> +1

<Curt> +1

<dgarijo> +1

<jun> +1

<ericstephan> +1

<zednik_> +1

<sandro> +1

<GK> +1

0 - not read

<tlebo> +1

<smiles> +1

<satya> 0 - not read it yet

<Luc> ACCEPTED: the three current documents PROV-DM, PROV-DM-CONSTRAINTS, and PROV-ASN should become the new Editor's draft.

<khalidbelhajjame> +1

<GK> (I haven't read the others, but I'm happy for them to be editor's drafts for now)

Luc: Last question, do we have the agreement we have reached, from F2F?

<pgroth> +q

Luc: can we resolve it?

pgroth: suggest there are still editorial adddress to address first
... before we can say we have achived the goal

<tlebo> +q to say that I think WD4 handles "conceptual versus technical" but not "scruffy versus proper"

pgroth: as GK pointed out, we can discuss that once it's public

GK: resolution that matters is when we release it (?)

<Zakim> GK, you wanted to say I think the resolution that matters is when we agree to release a new PWD

Tim: As Jun gave her feedback, I realised that clarity is conceptual vs. technical. That transition path that we promised, that Ivan passes to distinguish .. (?)
... others agree with that?

Luc: good point, time to talk about process
... as we agree they will become editors working drafts, we can raise issues in the tracker

<GK> @tim do you mean what we've been calling "scruffy/precise" transition?

Luc: and a point like that, Tim, can be raised as an issue
... and then debate -> resolve it

<GK> ... if so, then I assume we'll work on that

<Luc> decide whether ISSUE-145, ISSUE-183, ISSUE-215, ISSUE-225 and ISSUE-234 (all relating to identifiers) can be closed

<tlebo> @gk ??

Luc: another point addressed from review - can issues relating to identifiers be closed?
... perhaps do that offline due to time constraints. I propose to close it, and those who raise it will answer

<GK> @tim When you talked about conceptual vs technical, I meant.

pgroth: set a time limit in the email
... for responses

PROV-O Ontology: Reviewer feedback

Luc: completes PROV-DM

<tlebo> @GK, I think WD4 addresses conceptual versus technical, but DOES NOT handle scruffy versus proper.

Luc: feedback - skip myself for now.. Paolo?

<pgroth> @tlebo - i would disagree

Luc: number of issues.. good alignment, simplified.. compliant, if it was leading to natural RDF

Paolo: first 2-3 points.. short summary: right direction

<pgroth> @tlebo as curt said scruffy means you didn't read part II

Paolo: started looking at it on Monday. Many things I would have pointed out has already been addressed
... others in my email might have been addressed already
... alignment with hierarchy, devil is in the details (?)
... not seen any reply to my comment yet.

<tlebo> (oh goodness, perhaps I missed the second two parts!)

Paolo: good alignment

<tlebo> @all, sorry...

Eric?

tlebo: but that's a vlid point that it's easy to miss the other parts :)

ericstephan: missed deadline.. still time to comment?

Luc: all comments useful.. but lots of traffic to catch up :)
... now moving target.. wait a few days?

stephenc?

stephenc: Looked in Protege, looking at ProvRDF mapping
... which makes sense
... structure of classes, hierarchy of classes and properties make sense
... adressing question of naturalness.. I was interested in if you can say simple things simply

<pgroth> yes

stephenc: like are we specifically allowed to use binary relationships without the Involvements

(yes)

<Paolo> @stian: s/devil is in the details/details are in my mail :-)

stephenc: to use it in OPMV style, use the simple relations for simple things

<tlebo> @stephenc, yes, the binary relations can be used on their own.

stephenc: lots of stuff with characeristics of properties, transitivity, symmetry, etc.
... would be nice to see lots of the properties tied to gether by property definitions (?)

<tlebo> @stephenc, the binary properties are defined, what suggested that you couldn't just use them?

stian: what did you mean?

stephenc: for instance used property can be thought of as used qualified involvement
... if you could use properties from the qualified involvement to infer the used property
... and then what informed by, used and qualified

<satya> @Stephenc: good point, we need to model them as rules

<satya> @Stian: +1 (separate from owl ontology)

<khalidbelhajjame> @Stephane, I think inference will the model more complex, woudn't it?

Stian: We have kept various things like that out to keep it in OWL-RL, but those kind of inference rules could certainly be tacked on as additional OWL file or rules

stephenc: at one point I noticed that the way that the properties are defined, you can use the same proeprties

<pgroth> aren't property chains in owl-rl?

stephenc: like the qualified.. that makes that more difficult

<khalidbelhajjame> My hope is that at a later stage when both direct binary properties and the classes of involvement are stable, we can have a light prov-o with only the binary properties

Luc: (?) did you go through OWL?

?: The ProvRDF mapping file was useful, loaded OWL in protege, but did not have time to check out everything

<tlebo> stephenc: "inverted" prov:qualified property will make property chains less direct to create.

^^Curt

simon: Feedback.. before ProvRDF mapping, my feedback was what I know how to use it for the primer
... My comments are small, it seems to make sense, what are ranges of some properties like had Location, and why they are part of model at all

seemed separated from ontology

Paul?

pgroth: going in right direction
... of being consistent, and given constructs for all DM records
... still some issues that are being, need to be addressed. In particular conversations around how we distinguish what is part of the serialisation
... like we can do it in OWL-RL.. DM.. what is in serialisation.. what is the model of the DM
... but a good step in right direction
... in reflecting DM

Luc: any other comments before I give my feedback?

pgroth to chair if discussion starts (!)

Luc: Key question was if the ontology is aligned with DM
... did not go through all the relations, but focused on activities, entities, derivation, usage, association

generation

<pgroth> @tlebo: http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.02.23

Luc: I see as core of model - if that is solved properly

<pgroth> in the agenda

Luc: my intuition at this stage is that what we can express in DM can be encoded in the ontology, as explaine by ProvRDF mapping

<tlebo> @pgroth, thanks.

Luc: I've implemented part of it
... working fine
... issues that are raised, number of things you can express in ontology that are not in DM
... paolo mentioned something, like time information that can be attached to instances in RDF where there is no DM equivalent
... another is that PROVO provides a structure for the concepts of DM, that's nice
... properties such as qualified, involved, and some classes, prov:Involvement etc
... but it means you can use these classes and properties - all part of the structure - and no DM equivalent
... what are we trying to achieve?
... interoperability concern
... if we think about that, then we need to express what is in RDF to map it to other technologies
... they may not have all the same notions
... if it is not part of data model
... every mapping to a technology would include nice features
... if you do an XML mapping then you could also do interesting XML encoding tricks
... I've seen an object-oriented style mapping with abstract classes
... which would make sense there
... what we need to do is to distinguish core of DM, and what is not core
... what are the nice features.. mapping specific
... at the moment, the ontology has both
... mixed together
... suggested earlier on how we could address these by separating PROV-DM specific notions from the nice features from OWL
... and let users decide

<GK> Surely, the primary point of interop is to be able to map the valid DM features ; making errors impossible is a different goal, IMO

Stian: open world assumption, etc - not sure if it would be possible to split

Luc: you can send that an entity qualified usage of another entity
... that is allowed by ontology now
... which is not part of DM

(OK, that's a fair point)

<satya> @GK, Stian: +1 (all languages have additional features and adding constraints for error checking is different)

<Zakim> GK, you wanted to suggest the primary point of interop is to be able to map the valid DM features ; making errors impossible is a different goal

GK: Two issues.. primary interoperability goal is to exchange between technologies
... Not sure if comments here prevent that
... ANother goal - not invalid - but how can you limit the things you can express so that everything in one technology can be mapped to another
... for instance if one can limit what the RDF permitted/conformant with OWL, then fine.. but might get too hung up in this when it's not really fundamental for interoperability

<Zakim> tlebo, you wanted to ask for summary of the sudden RL requirement

<khalidbelhajjame> @Tim, yes Luc sent an email

(yes - OWL is not meant to be used for restrictions - but possibilities)

<dgarijo> i think Ivan proposed it

tlebo: when was OWL-RL really agreed?

pgroth: Ivan mentioned that OWL-RL thought that this was encouragable
... to increase adoption
... and there was some census.. and now it's there
... we all along said that the ontology , should be 'lightweight' - we didn't define that earlier, at F2F it came out that use of OWL-RL would be that

<GK> AIUI, OWL-RL is a subset that is easily implemented in query systems

tlebo: will not raise my concerns here

pgroth: Luc - so are you saying that.. the current ontology does not give good alignment with WD3?
... a valid consern, but that's my question

<pgroth> q0

<jcheney> Luc: Can you be precise about what "not aligned" means?

Luc: I believe anything in DM can be encoded in PROV-O

<tlebo> Luc's concerns can be addressed with non-RL OWL constructs. We're getting our hands tied.

<sandro> tlebo, I think Ivan and/or I would be happy to talk about the RL issue in email.

Luc: ontology allows many other things to expressed.. like my entity-with-qualified-usage-using-another-entity
... that's too much to me
... allowing things to be expressed that should not be expressible
... too permittive

<tlebo> @luc, "permissive" is fixed with axioms that RL doens't allow.

@tlebo +1

<satya> @tim +1

Luc: like the patterns.. but try to separate what is really DM compatible vs what is nice patterns

pgroth: what does that mean in terms of process

<dgarijo> @tim: we could adress the problem by subtyping qualified..

Luc: notion of time is crucial to data model
... the reason why we've associated time to specific concepts if because we think there's the notion of event.. and a kind of temporal mapping with events
... notions such as assocation, responsibility.. where we did not include time
... nobody came up with a temporal mapping that made sene
... but if ontology allows time to be associated with almost anything, what does it mean to temporal constraints?

(my take: about the same as if there was random attributes like ex:started="yesterday"]

<khalidbelhajjame> @prov-o team, luc in his email already suggested one solution that looks fine to me, I didn't have an issue with it.

Luc: but that needs to be addressed

Satya?

Satya: To clarify.. adding time to every construct, how does it prvent it from validating according to DM constriants?

sorry I can't scribe

lost battery

<scribe> NEW SCRIBE please

<GK> Paul's question: does this prevent us going forward with this document?

<dgarijo> I'll scribe

<dgarijo> luc: we need to reflect that in the data model. Nobody has done that

<dgarijo> ... I'm not saying that DM is complete, but it is not aligned.

<dgarijo> satya: we have similar issues with location

<dgarijo> ... the domain is everything

<tlebo> OWL is not about preventing people from asserting silly things, it's about adding more useful things based on what was said.

<GK> I don't think DM should be changed to match constraints expressible in OWL. TAILS WAGGING DOGS COME TO MIND

<GK> @stephan +1

<dgarijo> zednik: don't understand why do we have a restriction on silly statements

<dgarijo> ... if someone wants to make it, ok, but it's not our concern

<khalidbelhajjame> Yes, this issue has already been raised by Daniel

<dgarijo> pgroth: need to identify which parts of prov-o are more expressive than DM and add a text explaining how not to use

<satya> @pgroth: is that part of the best practices?

<GK> @pgroth +1

<dgarijo> I think it makes sense

<dgarijo> @pgroth: +1

@dgarijo I happy if you can continue scribing as I'm back on old-style landline

<dgarijo> luc: what is the concrete proposal for the prov-o team

<dgarijo> @stain: no prob

<dgarijo> ...?

<pgroth> Proposed: current owl file reflects wd3, the prov-o team should mark where the prov-o allows more expressiveness than the dm and should come up with proposals to see if it's possible or doable to address these constraints

<dgarijo> :)

<dgarijo> jcheney: I don't get the problem: what is the property of prov-o that it shouldn't have?

<GK> @paul That's two parts. I fully support 1st part; 2nd part I half support.

<Curt> If someone writes bad prov-o, it would prevent interoperability with other prov formats/languages/etc.

<Luc> http://www.w3.org/TR/2012/WD-prov-dm-20120202/#record-relation

<Curt> I think that's ok ;-) GIGO

@Curt: so perhaps the question is - how can you detect bad PROV-O

@Curt: ie. a set of rules or OWL-Full constraints

<dgarijo> luc: I wrote an email (it's on the agenda)

<satya> @Curt, Stian: rules

<dgarijo> ... usage can be used between 2 entities, for instance

<jun> @curt, or examples?

<dgarijo> luc: the solution is go to the email and discuss it.

<dgarijo> pgroth: other solution would be to write: DON'T DO THAT in the scpec document

<jun> rules and constraints would require an implementation of validator. and would it scale?

<zednik_> we can use restrictions that put us out of OWL-RL, or annotations in the ontology to guide usage

{ ?x prov:qualified ?usage . ?usage a prov:Usage; prov:entity ?y } =? { ?x a prov:Activity; prov:used ?y . ?y a prov:Entity }

<dgarijo> jcheney: missinterpreting what Luc said.

<GK> @jun it wouldn't be mandatory to actually *use* rules and validator

exactly

people are even allowed to use the OWL ontology without knowing much about OWL

<dgarijo> luc: we have to be precise in the alignement. We should be able to express DM in prov-o, but also prov-o should not be more expressive than DM

use it as an RDFS vocabulary

<tlebo> @jcheney, http://www.w3.org/2011/prov/track/issues/262 ?

<jun> @gk, ack. gotcha

<dgarijo> jcheney: 1)people have been pointing that fixes to your problem would break owl-rl

<Curt> A 'prov validator' could go beyond the simple expression of prov-o

<satya> 1. There will always be issue translating from OWL to XML or other languages (not everything can be "carried" over)

<dgarijo> ... 2) If we don't know what the alignement prop is then how are we going to align it?

<Luc> That' s how I suggested we can address the issue http://lists.w3.org/Archives/Public/public-prov-wg/2012Feb/0414.html

<pgroth> q

<satya> 2. Adding error checking rules will (I think) be out of RL profile

<GK> I think anything that is satisfiable in the formal semantics should be round-trippable without loss of information

<dgarijo> luc: I don't have a formal ..?..

<dgarijo> ... an entity having a qualified usage of an entitity is not the intention of DM

<dgarijo> ... I made a suggestion on the email

<satya> 3. Adding inference rules (as Stephenc suggested) will be definitely require rules (most probably in RIF)

<dgarijo> ... I am concerned about the interoperability issues

<Paolo> have to go now, apologies

<GK> My definition of interoperability above: anything that is satisfiable in the formal semantics should be round-trippable without loss of information

<dgarijo> ... maybe I've a stronger interpretation than others. Maybe we need that definition

@GK but I agree with Luc in the sense that the OWL should guide you towards interoperability, and not lure you directly into non-translatable things

<dgarijo> pgroth: the ontology reflects wd3, but it has more stuff

<stephenc> @satya I was only suggesting using owl:propertyChainAxiom, which is in OWL-RL

<dgarijo> ... that shouldn't be a blocker

<GK> @stian: agree, but don't want to get hung up on this in the name of faux-interoperability

<dgarijo> ... how do we move forward? I'd like the ontology as is

<jcheney> @GK: there are two different interpretations: DM -> Owl -> DM (which I think "works" now) and OWL -> DM -> OWL (which I don't think "works" but I'm not sure it is what Luc means).

<dgarijo> ... but we could raise issues

@GK: agreed. Restrictions can be tacked on.. and getting EVERYTHING restricted so it's not possible to express something that does not map to DM would be very hard.

<dgarijo> luc: agrees

<dgarijo> luc: what Tim thinks about this?

jcheney: no, but that would not work unless DM had a complete 'any RDF'-node everywhere - which perhaps was the idea with the 'attribs' - but it is not enough

<dgarijo> tlebo: james just said what I wanted to say

<GK> @jcheney if formal semantics reflects/drives DM constraints, then surely any OWL that is satusfiable in formal semantics *is* riound-trippable?

<dgarijo> ---a lot of typiing noise!!--

<jcheney> @GK: not sure yet...

<dgarijo> luc: there have been some recent changes

<dgarijo> @GK: thanks!

the ontology has always allowed even :entity1 prov:used :entity2 as :Entity and :Agent was not stated as disjoint (that's out of RL)

<dgarijo> tlebo: removing all the subprops of qualified was a move to simplify the model

eh// entity and activity

<dgarijo> @stian: you are actually right..

<dgarijo> tlebo: a lot of different kinds of requirements

<dgarijo> ... we still don't have a corpus of examples that address these concerns

<dgarijo> ... the way of not forgetting about this issues is to have examples in our repository

<dgarijo> luc: it is a very good idea

<satya> @tlebo: +1 (saves prov-o from trying to satisfy moving requirements)

<GK> Test cases are good.

<dgarijo> pgroth: I don't understand what the conclusion here is

<dgarijo> ... right now it is raised as an issue, but I don't know where are we going

<dgarijo> luc: I invite prov-o team to review the feedback

<dgarijo> ... it will be good to see what the response is

<GK> Question is "decide whether the ontology offers a good alignment with prov-dm wd3" - but what does this mean? What really matters is can we proceed with this?

<dgarijo> ... and analyze whter it can be modeled or just warn in the html spec

@Luc

<dgarijo> pgroth: what's next for that team?

<tlebo> (just blacked out for a minute)

<dgarijo> ... can they start working on the doc?

<dgarijo> ... solve all the issues of the ontology first?

<dgarijo> +q

<khalidbelhajjame> I would prefer the option of focusing on fixing the lain issues of the ontology before trying to revise the HTML documentation

<Luc> what do other reviewers think?

<dgarijo> stain: agrees with daniel. Document what it's obvious, and not document the parts with issues

<tlebo> the HTML needs to stop being postponed.

<GK> @paul +1

<dgarijo> pgroth: wouldn't want to get hung up on this point

<jun> as long as the parts with issues are kind of self-contained, I agree with paul and daniel

<dgarijo> ... we shoud decide on whether the issue can be addressed reasonably or not

<dgarijo> ... issues 64, 262..?

<dgarijo> ah ok

<Luc> 253, 262, 263

<GK> Alternative definition of interop: any RDF that corresponds to a valid DM expression can be round-tripped without loss of information. I think that covers RDD-ASN-RDF and ASN-RDF-ASN.

<pgroth> proposed: prov-o team look at the issue 253, 262, 263, etc and see if it can be addressed but this should not hold the group back

<dgarijo> +1

<GK> +1

<khalidbelhajjame> +1

<satya> +1

<ericstephan> +1

<zednik_> +1

<Luc> @GK, yes, but can we determine, in rdf, what is a valid translated dm expression?

<tlebo> bye bye!

<dgarijo> pgroth: bye

bye!

<ericstephan> see ya!

<khalidbelhajjame> bye

<GK> @luc: I think so, but maybe not using OWL

<pgroth> are you doing the minutes luc?

<GK> Bye.

<Luc> It would be good if it could be mechanical!

<GK> @luc: it would be good, but not a disaster if not

<pgroth> trackbot, end telcon

- DRAFT -

Provenance Working Group Teleconference

23 Feb 2012

Attendees

Contents

admin

Review of actions

PROV-DM Simplification: Reviewer feedback

PROV-O Ontology: Reviewer feedback

Summary of Action Items

Scribe.perl diagnostic output