Provenance Working Group Teleconference -- 13 Oct 2011

<trackbot> Date: 13 October 2011

<Luc> Scribe: Daniel Garijo

Admin

<Luc> @dgarijo, hi daniel, everything is set up for you

Hi Luc, thanks a lot!

<pgroth_> Zkaim, +[IPCaller] is me

<stain> finally recognized by the conference bridge!

<stain> http://www.w3.org/2011/prov/wiki/Meetings:Telecon2011.10.13

<Paolo> zakim ??P38 is me

Luc: look at the name of the formal model
... review of the agenda. Any other issues?

<Luc> PROPOSED: to accept the minutes of Oct 06 telecon

<satya> +1

<khalidbelhajjame> +1

<Curt> +1

<SamCoppens> +1

<stain> +1

<Paolo> +1

<Luc> ACCEPTED: the minutes of Oct 06 telecon

Luc: no actions to review in the tracker.
... scribes still needed.

Name for 'ex-formal model' document

Luc: formal name was confusing
... what are the proposals of the ontology?

<stain> PROV Ontology (PROV-O)

stain: prov-ontology

<khalidbelhajjame> Prov-O

<pgroth_> +q

<stain> http://www.w3.org/2011/prov/wiki/NameSuggestions lists current proposals

Luc: any counterproposal?

paul: is prov-o general enough? (there could be other serializations not in owl).
... for instance if we use riff.

<satya> @Paul: yes, I think that is a good point - hence we did not include OWL or OWL2 in name

stian: should cover any technology in the future as well.

<pgroth_> ok seems fine to me

satya: ontology should be enough to adress the technology issues.

<stain> @Satya +1 (sound is not too bad, btw)

<khalidbelhajjame> For other technologies to which we want to map the provennce model, we can use PROV-ASN

<Luc> http://www.w3.org/TR/2011/WD-prov-o-20111013/ does look slightly strange.. is it 0 or O?

Luc: comments: is it 0 (zero) or O ?

<stain> or http://www.w3.org/TR/prov-o/ for the latest version

Luc: at some point we'll have to tackle named graphs.

<stain> Stian: Not particularly bothered by o/0 - just esthetics.

<pgroth_> +q

Luc: is it still right to talk about provenance ontology in that case?

paul: by using "ontology" we could include everything under that

<Luc> sandro ??

<Paolo> prov-OM?

Luc: if we decide this name, will we be able to change it in the future?
... prov-ONTO was also prpoposed

<pgroth_> +q

stian: an ontology is already a model.

stain: maybe the name should be similar to the paq document.

<satya> @Paolo PAM :)

<stain> @dgarijo - no, I meant that access-and-query is not a model, perhaps an architecture :) I meant that all 3 names won't match up with "model"

paul: people on the mailing list say provenance ontology. prov-o captures it well. Maybe prov-sw, but is not a big issue

<Luc> PROPOSED: to adopt 'prov-o' as the short name for the PROV Ontology

<satya> +1

<JimMcCusker> +1

<Curt> +1

<khalidbelhajjame> +1

<SamCoppens> +1

<Paolo> +1

<stain> +1

<ericstephan> +1

<smiles> +1

<Luc> ACCEPTED: to adopt 'prov-o' as the short name for the PROV Ontology

<pgroth_> +q

paul: how do we say it?

<satya> @Paul: like Bravo :)

<Curt> Rhymes with Bravo

PROV-DM FPWD

<satya> as Curt pointed out earlier ?

Luc: date of release.

<JimMcCusker> @Paul, I think that SKOS managed to avoid people knowing, I don't think it matters. I'm a skier, so I'll probably pronounce it like the city. :-)

paolo: it is done. Have you got the url for it?

<Paolo> http://dvcs.w3.org/hg/prov/raw-file/default/model/ProvenanceModel.html

<Paolo> http://www.w3.org/2011/prov/track/products/2

paolo: the situation is that the document reflect the discussion on some of the issues (not all)

<Luc> http://www.w3.org/TR/2011/WD-prov-dm-20111018/

paolo: some of these received enough discussion that has not been reflected in the document, but working on it. Other discussion (as in event), does not map directly to some of the issues.
... what is going to happen next?

Luc: try and prepare a timetable to see where we want to go in the next months
... bring proposals to vote for the next telecon, in order to be able to deliver the documents to the W3C on the dates

paolo: where is the input is going to come from?

<pgroth_> the paq doesn't get to say anything ;-)

paolo: already been interacting with other groups.

satya: regarding the issues that I was trying to raise: the role of constraints. We were trying to model the constraints in the ontology.
... suggest set of constraints as best practices.
... how should we consider this?

<satya> ok thanks

Luc: we should not address this rigtht now (not the right time)

<Luc> but this important, and we need to have this debate within the WG

paul: we have an official statement that weill go out to get responses from the people. Divulgation of the report.

Luc: connetion of the task force report.

eric: drafted some text with kai. Share some of it with Paul.

<khalidbelhajjame> would it be possible to share the text with everybody in the working group, just to see if ether are itmes that need to be discussed within the WG

eric: everyone is welcome to join the meeting

<kai> +1 (not on the phone today, just reading the minutes)

Towards prov-o document fpwd

<stain> http://dvcs.w3.org/hg/prov/raw-file/tip/ontology/Overview.html is current draft

satya: update on the document. 2 directions: 1) to meet some of the syntatic requirements (most of it done)

<khalidbelhajjame> Items to report from the Formal model (or ontology) group

<khalidbelhajjame> - Working on validating and making the html document compliant with the requirements.

<khalidbelhajjame> Satya identified the tasks need to be done for that purpose, and we assigned the list of tasks among ourselves.

<khalidbelhajjame> I believe that most of the issues has already been dealt with.

<khalidbelhajjame> - The OWL ontology was modified to be in line with the PROV-DM.

<khalidbelhajjame> - Similarily, the HTML document was updated to include classes and properties and be in line with the PROV-DM.

<khalidbelhajjame> - The Diagram representing the ontology was also updated. The diagrams representing the examples (Crime-File and workflow provenance) may still need to be modified to relect the new changes.

<khalidbelhajjame> - Satya is working on the constraints.

<khalidbelhajjame> - Any other thing Tim, Stian?

stian: added namespaces and fixed images (relative paths vs external links to the wiki).
... did xhtml vs html4. We can do either.

<satya> @Khalid: nice documentation!

satya: the second direction: we are aligned with the data model as much as possible.

<pgroth_> +q

Luc: estimated release date?

satya: last week we released one. We are continuing the process.
... right now can be considered as a released document.

paul: wondering on the status on putting all the concepts of the datamodel into the ontology model

<khalidbelhajjame> There is a section at the end of the document that identifies the concepts (relations) that are not considered in the ontology yet

<Luc> and vice-versa!

paul: you could give an idea of which concepts are NOT yet defined in the ontology.

@khalid, did we add the shortcuts too?

<khalidbelhajjame> @Daniel, not yet

@khalid, then we should :) I'll give it a try afterwards.

satya: will do that later

<khalidbelhajjame> @Daniel, yeas please go ahead, and we can discuss it in the next (Monday) telecon

Luc: very important point.

<Luc> This document is part of a set of specifications aiming to define the various aspects that are necessary to achieve the visition on inter-operable interchange of provenance information in heterogeneous environments such as the Web. This document defines the PROV-DM data model for provenance, accompanied with a notation to express instances of that data model for human consumption. Two other documents, to be released shortly, are: 1) a normative serialization of PRO

<Paolo> visition -> vision :-)

<stain> : 1) a normative serialization of PROV-DM in RDF, specified by means of a mapping to the OWL2 Web Ontology Language; 2) the mechanisms for accessing and querying provenance.

<Luc> , specified by means of a mapping to the OWL2 Web Ontology Language; 2) the mechanisms for accessing and querying provenance.

<stain> @Paolo, I like visitations :) Kind of invitation-visits

Luc: it should be clear to readers that all the concept on the data model document are taken in consideration in the ontology. And viceversa: if something is not on the datamodel, then it should be clear too

satya: the owl ontology is modelling of the data model, not necesarilly a mapping.
... Agrees with Luc, but some of the concepts may not be "encodable" in owl.
... we may not be able to map the model directly
... in to owl.

Luc: we are going to serialize the model in RDF, that is the message sent to the semantic community.
... Ivan was raising is that the kind of the ontology. Is owl-dl? is it simpler?

<stain> I think it's OWL-Lite at the moment

<pgroth_> could you talk about that in the document

<stain> but depending on how much of the constraints we need to describe it might increase to DL

satya: it should not be owl-full, since it is not decidable.

<Luc> is it OWL2-RL?

we haven't found anything that constraint us to anything more that owl dl

we haven't specified any specific constraints yet

<stain> it's closer to RDFS, plus owl:IrreflexiveProperty, etc

<satya> PROV-O has DL expressivity of ALR+

Luc: share the documents to get feedback from Sandro, Ivan and others.

<Luc> action on satya to ensure all terms of DM appear in prov-o document

<trackbot> Sorry, couldn't find user - on

paul: sum up of what has been decided in the telecon

<satya> @Paul: lightweight?

<Luc> action satya to ensure all terms of DM appear in prov-o document, to justify why other terms are introduced, and to explain lightweight nature of owl ontology

<trackbot> Created ACTION-40 - Ensure all terms of DM appear in prov-o document, to justify why other terms are introduced, and to explain lightweight nature of owl ontology [on Satya Sahoo - due 2011-10-20].

@Luc, thanks for recording it.

satya: what do we mean by lightweight?

<satya> @Paul :)

pgroth: (too much noise) when we say it's the rdf community means that not doing any crazy modelling of the ontology.

satya: stian has already demonstrated that we can use the ontology for some examples in different scenarios

<stain> agree

Interoperability

<stain> to not scare away people who have been bitten by the massive-owl-full-ontologies

<pgroth_> @stain - exactly

Luc: we will have to decide what we mean by interoperability

<Luc> http://www.w3.org/2011/prov/wiki/Interoperability

Luc: no answers yet
... document with some ideas to see if it makes sense.
... the document sumarizes the charter and certain aspects: define the data model independant of any toechnology and mapping to certain technologies.
... given this, what do we mean by interoperability? Ennumerated 3 types.
... it is not a complete set of types of interoperability. It is just to start the discussion.
... the first one: not loose any information when changing the representation.

second one: different representation with different forms of inference. What you inferr in one representation is desireable to be inferred in the other one.

scribe: thir one: similar to the provenance challenge objective: different provenance systems recording provenance in different formats: end to end interoperability through systems.
... terms in different vocaubaries: ns for the datamodel, ns for the ontology. Is it possible to reuse common urls?

<khalidbelhajjame> +q

scribe: this might not be complete yet, but enough to initiate the debate.

<stain> what is declared in http://www.w3.org/ns/prov-dm/ ? "role" and "type" only?

scribe: opinions?

satya: on what context are we discussing this?

Luc: to reach the level of recommendation, we should have at least 2 different approaches that are interoperable

satya: data integration. If we are using RDF for the formato of data integration, are we saying that we are going to define a new set of technologies, or reuse the existant ones?

khalid: prov language as an interchange format
... what is expresable with our language that is not with other languages
... and what are our limitations?

<satya> sorry I have to leave now, will try to follow up on this discussion over mails

<pgroth_> echo

<pgroth_> no

<stain> mute please

<stain> dgarijo: could you put it on IRC as well what you said?

dgarijo: we should not restrict ourselves to the vocabularies that model provenance , but also show examples with some tools (Taverna, wings, etc). We should look for potential clients that can adopt our model.

pgroth: what do we mean by interoperable between 2 serializations?
... how do we check that one serialization has used correctly the other one?
... build test cases

eric: how introperable are you?

<stain> convert testcase format a1->b1->a2->b2->a3 - compare a2 and a1 manually, automatically compare b2==b1 and a2==a3

<Paolo> the design of compliance test cases should be central to the interop effort

eric: seeing R model in owl.
... explain how we did the connections
... between 2 serializations.

Luc: the integration aspect is crucial, and potentially can be linked to some docs.

<pgroth_> we have a task force for this?

<pgroth_> no?

Luc: paolo, do you have an idea of how the compliance test cases should look like?

<khalidbelhajjame> Is it just me who think that talking about interoperability between different serializations of the same model is unusual?

<pgroth_> so maybe we should we start to engage this task force

<stain> I don't think interoperabable X would need to express anything we have in PROV?

paolo: how do you know that your serialization is compliant with the good one?

<JimMcCusker> gotta run to another call, TTYL

<khalidbelhajjame> +q

pgroth: task force on this?

<khalidbelhajjame> -q

Luc: engage the user task force in this.

paolo: is it realistic that we develop something to check compliance to the prov model?

<khalidbelhajjame> +q

<khalidbelhajjame> -q

<stain> could we have something similar to a validator?

Luc: not our job to give certification of compliance.

<stain> it could check the constraints etc

khalid: (too much noise) some ideas to check if a model is compliant?

@khalid: can you summarize, please? I coulnd't hear you well.

stian: check that something complies with a set of provenance assertions. It could be syntatic&semantic validation.

<khalidbelhajjame> I think it is not difficult to check the conformance to the model, but it is harder to check that the system use and use correctly the provenance

<Paolo> def need to go further than syntax validation

@khalid: thanks!

paul: those who have proposals of what interoperability should be, raise discussions on the mailing list

<Luc> or keep on editing the wiki page ...

Luc: action on Paul and Luc to engage the user task force.

<khalidbelhajjame> Ok

Luc: please edit the wiki page if you have further proposals.

<Luc> exit

<Luc> quit

Hi luc

<Luc> hi daniel, thanks for scribing, i'll deal with it from now

are you going to save the log?

thanks

see you !

<Luc> trackbot, end telcon

<trackbot> Sorry, Luc, I don't understand 'trackbot, end telcon '. Please refer to http://www.w3.org/2005/06/tracker/irc for help

<stain> trackbot, end telcon

- DRAFT -

Provenance Working Group Teleconference

13 Oct 2011

Attendees

Contents

Admin

Name for 'ex-formal model' document

PROV-DM FPWD

Towards prov-o document fpwd

Interoperability

Summary of Action Items

Scribe.perl diagnostic output