Provenance Working Group Teleconference

01 Sep 2011


See also: IRC log


Luc Moreau
stain, GK, Stian


<trackbot> Date: 01 September 2011

<Luc> Scribe: stain

<Luc> @stain, are you scribing?

<Luc> we don't seem to have a scribe

<Luc> Scribe: GK

<stain> Luc: sorry I was late

<Luc> stain: do you still want to scribe, we are starting ...

<stain> skype was not playing along

<stain> Luc: Requirements for main draft

<stain> Luc: Talk about provenance ontology

<stain> Luc: suggests to drop document from agenda

<scribe> Agenda: accept minutes of Aug 25 telecon

<Luc> PROPOSED to accept the minutes of Aug 25 telecon

<Luc> http://www.w3.org/2011/prov/meeting/2011-08-25

<satya> +1

<Curt> +1

<Paolo> (wasn't there)

<dcorsar> +1

0 (not present)

<stain> 0 (not present)

<Edoardo> +1

<smiles> +1

<StephenCresswell> +1

<Luc> ACCEPTED: the minutes of Aug 25 telecon

Minutes accepted

<scribe> AGenda: review actions

Named graphs requirements

No outstanding actions

<scribe> Agenda: Named graph requirements

<stain> Luc: The RDF working group would like to have a telcon to hear our requirements

Luc: RDF WG would like teleconference to understand requirements.

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

<stain> Luc: A wiki pake made by Satya

@Stian - are you taking over

<satya> @Luc: Is there a tentative date for the telcon?

<stain> @GK sure - if you mute your keyboard :)

<scribe> Done.

<scribe> Scribe: Stian

<stain> Luc: 2011-09-15 The 15th of September as tentative date - after the normal telcon - extra 45 mins

<stain> Luc: But not yet decided

<satya> me and Paul

<stain> Luc: Who submitted the requirements of the wiki? Could authors indicate?

<stain> Luc: Any other requirements? Simon?

<stain> Simon: All there

<stain> Satya: Wanted to add more points before telcon

<stain> Luc: Also had some ideas - need to check if they are captured

<stain> GK: By email - nature of provenance. Two possible roles for named graphs - which might be quite specific

<stain> GK: 1) Handling of accounts

<stain> GK: 2) Handling of contextual assertions of provenance

<stain> GK: See my response in email response to Luc

<stain> GK: Need a way to encapsulate provenance statements to relate to a context

<stain> GK: Suggest to not discuss this in this telcon as it can be complex

<jorn> Zakim: ??p42 is me

<stain> Luc: Could GK add this as a potential requirement on the wiki page?

<stain> ACTION: GK to Add potential contextual named raphs requirements to http://www.w3.org/2011/prov/wiki/ProvenanceRDFNamedGraph [recorded in http://www.w3.org/2011/09/01-prov-minutes.html#action01]

<trackbot> Created ACTION-37 - Add potential contextual named raphs requirements to http://www.w3.org/2011/prov/wiki/ProvenanceRDFNamedGraph [on Graham Klyne - due 2011-09-08].

Name for the standard

<stain> Luc: First draft to be released end of month - need a name for the model/language/etc. Proposals http://www.w3.org/2011/prov/wiki/NameSuggestions

<stain> Luc: Straw poll here - but want to hear what people think about names

<satya> @Luc WE are counting PIF twice?

<Luc> 1,1,1

<stain> Luc: Particpants of the call have 3 votes - you can vote 3 on same name, 1 vote on 3 names, 3+1, etc

<stain> Satya: Worried about double-counting PIF (#15, #16)

<stain> Satya: Suggest counting 15+16 as one

<stain> Luc: 15+16 now merged on wiki

<dgarijo> +3 to PIL

<stain> Luc: Indicate which ones you are voting for now

<smiles> 11, 11, 15

<satya> 15, 15, 15

<Paolo> 3 14 15

<Lena> 3, 10, 15

<Curt> 5 15 14

<dcorsar> 3, 5, 14

12, 13, 15

<stain> 5, 10, 12

<dgarijo> (5, 5, 5)

<sandro> 10 15

<Edoardo_> 3, 5, 11

<stain> Luc: We'll count the votes offline and send an email and hope to progress from there

<stain> echo, echo

<zednik> 5, 5, 15

Primer Document

<stain> Luc: Discussed this 4 weeks ago, and said not to do a primer at that stage. Paolo and Luc as editors of model documents tries to illustrate the model - but also to specify it. It's difficult to do both in same document.

<stain> Luc: GK commented that this is not so useful - it's time to get on with a Primer document. Simon has expressed interest on worker on primer.

<stain> Luc: Would you want to work on a primer - and what do you expect?

<stain> GK: I commented that a 50k feet view would belong in the model. Don't seem to be completely clear in consensus in what model contains, so uncertain about doing a primer now while model still in flux.

<satya> +1 for GK's point

<stain> GK: Also said that example was not useful - the complexity was such that it was as hard to understand to example as the message

<Zakim> GK, you wanted to say that I think a 50,000 foot view belongs in the model, not a separate primer

<stain> GK: Perhaps that example would fit better in a primer - but still seed need for a overview in the model

<stain> smiles: example sounds good to include in the primer

<stain> smiles: more clarity step by step - say why things are done how they are. Might be reasonable to start with a simpler example

<stain> smiles: high-level view on model, agree with GK. High-level in model doc, but also in the primer in more non-normative terms.

<stain> smiles: A high-level description in the model document might easily always be normative - easier to suggest how to interpret model in the primer

<stain> satya: Agree with previous, primer has 3 functions. 1. Simple example. How would model elements be used in non-normative description.

<stain> Satya: 2: How would this be modelled in OWL/RDF, bits of ontology doc. 3: How would it be accessed - elements of query document.

<stain> Satya: Then give overall overview of how to handle provenance information

<stain> Satya: Now is not the optimal point for working on primer - look at it in the end of september when draft is published and discussed issues have settled

<Paolo> +1 for starting after initial model doc has been released

<Zakim> GK, you wanted to say I think the RDF modelling should be NORMATIVE, else we don't have a usable spec

<stain> GK: The representation in RDF should be a normative output - assumed that abstract model -> RDF would become part of model document.

<stain> Luc: Answer - No. RDF representation not in model document.

<stain> Luc: It would be in the formalisation document led by (?)

<Paolo> RDF repr should be somewhere but not in the conceptual model doc

<stain> @Paolo - yes - it should be formal, but not neccessarily part of the conceptual model

<stain> Luc: Illustrations can be done in RDF and/or the abstract .. - but not by end of September

<stain> GK: If this is to be useful on the web we need something to interoperate between application, and at least one normative format like RDF would be required

<stain> Luc: The normative spec will be included in the doc made by Satya

<Paolo> @GK mapping to RDF /is/ normative but in the ontology doc

<stain> Satya: The Ontology is the normative representation of the model. The illustrative RDF should corresponding to the normative OWL

<stain> Satya: Illustrated examples would be by the normative RDF format

I'm Ok with the normative mappingt to RD being in the ontology doc

<stain> Luc: Is the feeling to wait until end of September?

<Lena> me + stephan have defined the primer to be on our task force

<stain> Luc: Example is to have an example explained - ultimately a primer view

<Lena> but we need the model to be described first

I think when to start the primer depends on whoever wants to do the work :)

<stain> Satya: Could be useful as Smiles would work on primer, if he participates more with the other 3 groups and identify content that can be migrated to the primer later

<Lena> i agree with graham that it would be a wild goose hunt to produce a primer while the model is a shifting target

<stain> Satya: Not a separate wiki page!

<stain> Satya: Do as comments as part of provenance.. query.. task force wiki page

If start primer now, I think it should start as an proper draft document. +1 to not creating yet another wiki page.

<stain> Stian: If we don't have a primer, will there for the initial draft be an overview document that shows a quick introduction of what the model/ontology is, etc.

<stain> Paolo: If someone works on primer now, should shadow the other work

<stain> Paolo: Should start work on a complete example - iterative process

<stain> @GK +1

<stain> Paolo: Would inform the primer

<stain> Paolo: See mutual benefits from Smiles and others shadowing

<stain> Luc: Paolo suggests working on complete example - a new example to design, or data journalism example?

<stain> Paolo: Data J example not used in conceptual model docs

I'd say several examples: simple to complex, chosen to illusrate and/or test different points.

<stain> Paolo: If it is a good example, but no constraints

<stain> @GK, agree - show different bits instead of a massive example to learn first

<satya> @Paolo: :)

<stain> Paolo: Should anyway be a different example from the model document

<stain> smiles: an example section in the concept model document and formal document, and in access document - and then someone to edit those sections so that they are explained. Would that make sense?

<stain> Paolo: Not quite - something complementary

<stain> Paolo: That the model is sound and explain it, but a different model. Primer should not be part of the other documents.

<stain> smiles: Not the impression I got from others - they suggest it was too early, just wanted better examples in the normative documents

<stain> smiles: not much in either way

<stain> smiles: someone can work on the example within the context of the existing docs

<stain> Luc: Then to replace the example section with something different? Don't want two examples in conceptual model doc

<stain> smiles: expect several examples in the primer. add more text to existing examples to explain and clarify at this stage

+1 independent choice of what goes in model document

<stain> Luc: have a problem with File example does not cover all the concepts

<stain> Luc: Would be useful with an example that highlights all concepts

<stain> smiles: raise issues in that regard - edit the example

Is it appropriate/necessary for the model document to have examples covering *all* aspects?

<stain> satya: in the primer document with should have one example, and use this in all the documents.

<stain> satya: would DataJournalism example be better, Luc?

If model document is formal description, that should cover the essential content, methinks

<stain> Luc: DataJ example is fairly long, does not expose all concepts either

<stain> satya: Try to create snapshot examples for each model, in provenance ontology group we discussed this. Then do bottom-up approach for each element - then consolidate as one big example

<Lena> +1 for bottom-up!

<stain> Luc: Worried about changing at this late stage before going first public draft

<stain> satya: keep file scenario, but what we have in mind, we write in down in the conceptual model, as it evolves, keep append it to the top scenario

<stain> zednik: what are we compiling all into one example?

<stain> zednik: may have a complicated, silly example if we merge everything into one

<Lena> +1 for stephan's comment

<stain> @zednik +1

<satya> @Stephan - good point!

@zednik +1

<stain> Luc: So we're not quite clear yet - think a bit more and come up with a concrete proposal for next telcon

<Paolo> @zednik +1 -- conceptual model already has two separate scenarios

<Paolo> (file editing, Royal Society)

<stain> ACTION: smiles to Make proposal on how to proceed on primer material [recorded in http://www.w3.org/2011/09/01-prov-minutes.html#action02]

<trackbot> Created ACTION-38 - Make proposal on how to proceed on primer material [on Simon Miles - due 2011-09-08].

Adopting naming conventions

<stain> Luc: Did not reach consensus last week on past vs present tense

<stain> Luc: Look at what we said on the edges - some confusion last week. Proposal in the agenda

<dgarijo> can we post the proposal please?

<Luc> proposed: Edge labels contain a verb

<stain> http://www.w3.org/2011/prov/wiki/Meetings:Telecon2011.09.01#Adopting_naming_conventions

<Luc> isUsedBy, isControlledBy, isDerivedFrom, hasParticipants

<dgarijo> +q

<stain> Luc: Example of what form of verbs we would have

<stain> dgarijo: agree on having verb in the edge - but would strongly disagree on 'isUsedBy'

<stain> dgarijo: an artifact is used by a process and produced by a process

<stain> Luc: not suggesting these actual labels - but the general principle

<stain> Luc: Not proposing domain/range direction now - just that the term contains a verb

This is an area where I think a 50,000 foot view would help

<stain> Luc: Model document lists all labels.. but we're trying to adopt a convention if verb is explicit

To see all the names related as a composite structure ... see if they make sense together

<stain> Stian: Not sure if we'll get too much of isSomethingBy - could get a bit convoulted vs. 'something'

<stain> GK: This is where a bird eye view would help - a diagram showing classes and relationships - although simplified

<satya> @GK: you mean a visualization of the model?

<dgarijo> we have generated an overview of the ontology if that's what you mean, Graham: http://www.w3.org/2011/prov/wiki/images/a/ad/GlobalSchema.png

<stain> GK: Considering terms in isolation - then it's difficult as you loose the context in which they will work

<satya> @Daniel +1

<stain> Luc: in emails - Instead of isComplementOf - then say complement? hasParticipants -> participants

<stain> Luc: Satya reacted that this was not useful

<stain> Luc: Two conventions

<stain> Luc: what label conventions do we adopt

<Luc> proposed: Edge labels contain a verb

<dgarijo> +1

<satya> +!

<StephenCresswell> +1

<satya> +1

(Truth is, I don't feel strongly about the naming convention)

<Curt> +1

<Lena> +1


<smiles> 0

<stain> 0

<Paolo> +1 don't feel very strongly either

<Luc> ACCEPTED: Edge labels contain a verb

<zednik> +1

<Edoardo_> +1

<stain> Luc: If we have a verb - then we need to decide on the tense - like present or past

<dgarijo> +1 to PAST tense

<Edoardo_> +1 past

<Paolo> Past tense

<dgarijo> since provenance is describing things that have already happened

<satya> +1 for past tense

<stain> +1 to present

<Curt> The other issue was whether to always include the "is" or not.

<Curt> instead of "foo isControlledBy bar" just use "foo controlledBy bar" instead of "foo isControlledBy bar" just use "foo controlledBy bar" instead of "foo isControlledBy bar" just use "foo controlledBy bar"

(Next question: convention for passive or active voice? Slightly TIC)

<stain> Luc: Just voted to have a verb

<stain> Curt: "controlled by" has a verb

"controlled" is a verb

<stain> @GK right

<stain> ?: "used by" has a verb

<stain> GK: Sounds like passive or active voice

<stain> Luc: "Used" is a verb - "I used X" - but "X used by I"..

<stain> GK: Yes - it's a verb, but in passive voice

<stain> ?: Implicit or explicit verb

Sorry, that;'s the *active* voice

<stain> Luc: "used by" is not a verb - just "used\'

<stain> @GK, Perhaps passive is good as provenance is describing what went on

<stain> Luc: Unsure how to progress

Propose: editoprs come up with series of names and let the gropup comment

<satya> @Stian: Well it depends, I think active and passive are both useful

<Curt> +1 keep the explicit "is" verb, passive form, past tense

@smiles +1

<stain> ?: Just a case of people being unclear with what is a verb or not - the previous proposal was accepted - we're moving on to past tense or not

<Luc> proposed: To use past tense for verbs

<stain> Luc: Need to formulate a proposal

<dgarijo> the thing is that I would not be sure how to say that a process Used an artifact with "used by", since that means that an artifact is Used by a process

<smiles> +1

<stain> Luc: What is dgarijo suggesting..?

Example: used rather than uses ?

<stain> dgarijo: what would be the name of the edge of 'used' with the verb?

<stain> Luc: process "uses" an entity

<stain> Luc: But proposal for past tense means "process used entity"

used vs wasUsedBy - both past tense, but different directions

<stain> dgarijo: Why can't we use "control"

<stain> @GK - exactly

<stain> @GK and last week we talked about temporal directions

<stain> dgarijo: liked better "wasControlledBy" - but someone would ask if we say "used" then why not "controlled"

<stain> @GK the verb is good because it highlights exactly this direction

<Luc> proposed: To use past tense for verbs

<satya> +1

<smiles> +1

<dgarijo> +1 to past tense

<stain> +1

<StephenCresswell> +1

<Curt> +1


<zednik> +1

<Edoardo_> +1

<dcorsar> +1

<Paolo> +1

<Luc> accepted: To use past tense for verbs

<satya> agree

<stain> Luc: Implications for Satya and Paolo/Luc to update documents to use past tense and verbs

Provenance Ontology

<stain> @Luc shall I action that?

<dgarijo> @Satya, didn't you already put everything in past tense in the ontology?

<satya> http://www.w3.org/2011/prov/wiki/PIL_OWL_Ontology

<stain> Satya: No telcon on Monday, travelling/unavailable

<stain> Satya: made example of file scenario

<stain> satya: would like reviews and comments on this

<stain> satya: two objectives, define extension mechanism - how application can extend to model in their domain

<stain> satya: show how instances can be created using the ontology

<stain> satya: Pointed out that formal document includes RDF fragments showing encoding of the (?) scenario - welcome to have a look at this

<stain> satya: send comments as soon as possible

<dgarijo> @stain it is the crime file scenario

<JimMcCusker> +q

<stain> Luc: Would be useful to see the complete example encoded in RDF. In the document it's only shown a particular entity. Complete example?

<stain> satya: yes - it's at..

<satya> http://www.w3.org/2011/prov/wiki/PIL_OWL_Ontology#RDF_Graph_for_Crime_File_Scenario

<dgarijo> the image: http://www.w3.org/2011/prov/wiki/images/thumb/d/da/CrimeFileRDFGraph.png/900px-CrimeFileRDFGraph.png

<stain> satya: not suggesting to include the full RDF into the document

<stain> dgarijo: :(

<stain> JimMcCusker: IVPof and complement of seemsto have lost track of what is meant by those

<satya> @Stian: Daniel pointed to an auto generated diagram, we will re-structure that :)

<stain> JimMcCusker: might want to put together a page to start usecases of what this construct is adressing

<stain> @satya - I tried to zoom!

<stain> Luc: perhaps an model issue, not ontology issue

<stain> Luc: important to raise such issues on the tracker

<dgarijo> @stain: yes, maybe the rdf is better right now

<stain> Luc: discussion taking place - but difficult to understand what the problem is

<satya> @Stian: yes - we will create a separate diagram

<stain> Luc: Jim Mayers does not like isComplementOf as a label - but it's not raised as an issue yet - but he seems happy with the definition

<satya> I also have issue with the "complement" label

<stain> JimMcCusker: what was talked about with complement-of seems very difficult from IVP-of

<stain> Luc: Paolo and myself changed the definition to make them uniform

<stain> Luc: is-complement-of had to be revised to match entity's definition - but believe it's still the same spirit of original def

<stain> Luc: Please raise this as tracker issues

<stain> JimMcCusker: tasked to formalise this is-complement-of for the ontology group - it has been difficult to trace out

<stain> JimMcCusker: should ignore emails and look at the wiki?

<stain> Luc: no - the conceptual model document

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

<stain> Luc: but creating a set of complement-of usecases sounds good

<stain> Paolo: is illustrated with figure in model

<stain> http://dvcs.w3.org/hg/prov/raw-file/default/model/ProvenanceModel.html#concept-IVP-of

<stain> Paolo: feels responsible for this - so will engage with any discussion

<stain> Luc: Satya - what are the key issues now with ontology?

<stain> Luc: Last week you needed better understanding of model - is it better now?

<stain> satya: no - need to respond to email. What are top level concepts?

<stain> satya: we agreed ; two top-level concepts - but you suggest there could be more top level concepots

<stain> satya: perhaps Luc/Paolo to attend a telcon

<Paolo> agree that we (two groups) should interact

<stain> satya: fundamental - what are the top level concepts

<stain> Luc: Example of Role,.. what is not entity and process execution

<stain> (?)

<dgarijo> and we are currently discussing the modeling of n-ary relationship with the role-trick you proposed, Satya

<stain> satya: process is a type of entity - def is a continuent, location, etc.

<stain> satya: sounds quite broad definition, defined in many upper level ontologies

<stain> Luc: Should join next week's call

<stain> satya: Monday 12:00 eastern

<Paolo> ok

<stain> satya: could do separate if needed

<stain> Thank you!


<stain> Luc, what do I do now?

<stain> ok, thnx

<dgarijo> bye!

Summary of Action Items

[NEW] ACTION: GK to Add potential contextual named raphs requirements to http://www.w3.org/2011/prov/wiki/ProvenanceRDFNamedGraph [recorded in http://www.w3.org/2011/09/01-prov-minutes.html#action01]
[NEW] ACTION: smiles to Make proposal on how to proceed on primer material [recorded in http://www.w3.org/2011/09/01-prov-minutes.html#action02]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2011/09/01 16:11:49 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Found Scribe: stain
Found Scribe: GK
Inferring ScribeNick: GK
Found Scribe: Stian
Scribes: stain, GK, Stian

WARNING: No "Present: ... " found!
Possibly Present: Curt Edoardo Edoardo_ Example GK GK1 JimMcCusker Lena Luc MacTed P12 P17 P18 P21 P27 P29 P3 P32 P41 P42 P57 P75 P77 P78 Paolo Propose SamCoppens Simon StephenCresswell Stian Vinh aaaa aabb aacc aadd aaee aaff aagg aahh aajj aakk accepted dcorsar dgarijo joined jorn proposed prov sandro satya smiles stain trackbot zednik
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Agenda: http://www.w3.org/2011/prov/wiki/Meetings:Telecon2011.09.01
Found Date: 01 Sep 2011
Guessing minutes URL: http://www.w3.org/2011/09/01-prov-minutes.html
People with action items: gk smiles

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]