Provenance Working Group Teleconference

24 May 2012


See also: IRC log


jfuller, TomDN, Luc, +1.518.276.aabb, jun, tlebo, SamCoppens, Curt_Tilmes, dgarijo, khalidbelhajjame, GK1, +44.131.467.aacc, Sandro, +44.789.470.aadd
Paolo, Missier
Paul Groth
Curt Tilmes


<trackbot> Date: 24 May 2012

<pgroth> Scribe: Curt Tilmes

<dgarijo> hi all


<pgroth> http://www.w3.org/2011/prov/meeting/2012-05-17

<pgroth> proposed: approve Minutes of the May 17 2012 Telecon

<tlebo> +1

<dgarijo> +1

<khalidbelhajjame> +1


<jun> +1

<TomDN> +1

<smiles> +1

<SamCoppens> 0 (did not attend the meeting)

<pgroth> approved: Minutes of the May 17 2012 Telecon

pgroth: 2 open actions
... sandro to send working drafts

<pgroth> sandro?

pgroth: graham to review constraints. some emails flowing.

<Luc> there is an email that just arrived on the mailing list

pgroth: scribes, need them, please sign up

PROV-DM Restructuring

pgroth: past weeks, discussed several options
... graham made a proposal discussed last week
... chairs actioned to counter-propose

<GK> Sorry I'm late ... re actions, I just took a look at constraints. Basically looking good. Comments in email.

pgroth: many comments going around. Consensus proposal looking good, good approach
... core vs. extended has support on mailing list, try to confirm that today

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

luc: worked on text based on mailling list feedback
... revised section 2, now an overview with 3 subsections
... core, extended, organization and components
... 2.1 now has the diagram
... some work to go on content, ready for external review/feedback soon

<khalidbelhajjame> My only comment is that it looks good

<Luc> See figure: http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#prov-core-structures

pgroth: core structures in editors draft, are these what we want?

<tlebo> +q to ask about wasInformedBy

<GK> I think it's about right.

tlebo: prov-o also has wasInformedBy as a starting point, propose to add to DM core

<Zakim> tlebo, you wanted to ask about wasInformedBy

luc: that's the only difference between dm and prov-o now...

<jun> +1 to @tlebo

luc: a good reason to add wasInformedBy is that it makes sense in the core

<GK> I'm OK either way... I can see the case.

<smiles> Structure looks good; also makes sense for someone who's just read the primer, I think

luc: but it isn't really 'primitive' -- it is really a shortcut for other concepts in the core

<GK> Isn't the same true of derivation?

<TomDN> so is wasDerivedFrom, technically...

luc: I lean toward not including wasInformedBy in the core, but it is somewhat subjective

GK: Could go either way
... I see Tim's point

<tlebo> not sure that "primitiveness" should be the dominating consideration - the point is to give a conceptual basis to begin the story for any potential adopter.

<tlebo> +1 @GK !

GK: Is its primitiveness similar to wasDerivedFrom?

<Luc> used(a,e1) wasGeneratedBy(e2,a)

<Luc> used(a2,e) wasGeneratedBy(e,a1)

<Luc> wasInformedBy(a2,a1)

<pgroth> +q

<tlebo> Perhaps we should point out that this is the "core" of the DM, not the "core" of the constraints document. ("primitiveness" would be more appropriate in dm-constraints)

luc: ^used + wasGeneratedBy implies wasInformedBy

<Luc> used(a,e1) wasGeneratedBy(e2,a) does not imply wasDrivedFrom(e2,e1)

<GK> That's unexpected to me

luc: wasInformedBy is a real shortcut, but wasDerivedFrom isn't a shortcut, since you can't make such an implication

<GK> (Hunts in spec...)

pgroth: is having it in the core of value?

<jcheney> Definition currently given at: http://www.w3.org/TR/prov-constraints/#component-1--entities-and-activities

<tlebo> +q to restate that the purpose is to orient potential adopters, not to slip into the semantics among the constructs.

pgroth: people have asked about constructs to show certain relationships
... having it in the core helps understanding

<pgroth> activities aren't functions

<khalidbelhajjame> An activity does not have to use all inputs to produce a given output

GK: notion of influence is captured by derivation, you need an activity to create the new entity

<dgarijo> @Khalid: +1

GK: if we have got the consumption/generation, we have derivation

<khalidbelhajjame> Graham, you may be talking about traceability

<dgarijo> @Khalid: it may use e1 for a later generation than e2

GK: if an activity uses an entity, and generates another, then derivation is happening

tlebo: the purpose of 'core' is to orient new adopters

<Zakim> tlebo, you wanted to restate that the purpose is to orient potential adopters, not to slip into the semantics among the constructs.

<GK> @tlebo +1 that's more compelling :)

<dgarijo> @GK: Activities have an star time and an end time. That does not imply that all used entities influence all generated entities...

tlebo: other distinctions are secondary

<Luc> used(a,e1) wasGeneratedBy(e2,a) does not imply wasDerivedFrom(e2,e1)

<Luc> used(a,e1,t1) wasGeneratedBy(e2,a,t2) where t2<t1, so cannot imply wasDerivedFrom(e2,e1)

luc: dgarijo is right, you can't imply derivation in that way

<jun> @GK, I don't know whether what @Luc said is in the DM or not. But that notion was the part of the reason for the provenance community to have wasDerivedFrom

<GK> I see the point Luc makes, but I also agree @tlebos point

pgroth: Current core contents except for wasInformedBy is ok, is there strong objection to including it?

<jun> @GK, yes. I think the discussion was a bit off the original topic:)

<dgarijo> if it helps (as Tim suggested), then I think it should go in there..

luc: we haven't formally defined what is 'core', but it is misleading to consider wasInformedBy on the same level as the others
... but it is subjective, and we could include if the groups wants to
... I think it should not be there, since it is misleading

<jun> @Luc, misleading, but might still be helpful?

<pgroth> proposed: the core consists of entity, activity, agent, wasDerivedFrom, used, wasGeneratedBy, wasAttibutedTo, actedOnBehalf, wasAssociatedWith, wasInformedBy

<Luc> @jun, of course, i am not suggesting to drop it from the model, it is among the extended structures

<GK> +0.5 (happy either way, lean to this, but slightly)

<dgarijo> +1

<TomDN> +1

<tlebo> +1

<jcheney> +1

<smiles> +1 (ambivalent about wasInformedBy)

<stephenc> +1

+0.5 (ambivalent as well)

<zednik> +0.5

<SamCoppens> +1

<jun> +0.5 (because time was all excluded, but I am fine with it)

<khalidbelhajjame> +0.5

<pgroth> accepted: the core consists of entity, activity, agent, wasDerivedFrom, used, wasGeneratedBy, wasAttibutedTo, actedOnBehalf, wasAssociatedWith, wasInformedBy

pgroth: ambivalence plus support, so that seems the way to go

luc: should consider the choice

<pgroth> proposed: the core consists of entity, activity, agent, wasDerivedFrom, used, wasGeneratedBy, wasAttibutedTo, actedOnBehalf, wasAssociatedWith

<GK> +0

<dgarijo> +1 as well

<khalidbelhajjame> +1

<TomDN> +0

<smiles> +1


<jcheney> +1 happy either way

<SamCoppens> +1

<tlebo> -1

<Luc> q|+

tlebo: what does this mean for startedAtTime/endedAtTime for its starting point

<Zakim> tlebo, you wanted to ask if prov-o must remove "startedAtTime" and "endedAtTime" because of this resolution.

pgroth: prov-o should remove those, so core is consistent

<GK> I assumed that each document could make it's own evaluation

<jun> +0 fine either way

luc: no, prov-dm core doesn't talk about some things like time

<dgarijo> @Tim: when you say remove you mean to put it in expanded terms, right?

luc: we could, but I don't think prov-o would have to remove those from its starting point

<tlebo> @dgarijo, yes. Moving them to expanded.

<tlebo> thanks!

<dgarijo> @tlebo: ok, thx

pgroth: ok, you could leave them in the starting points

<GK> The question is: wasInformedBy core or not?

pgroth: looking at votes, leans to add wasInformedBy to core

<dgarijo> @GK: according to the votes it should be included in the core.

<pgroth> proposed: is wasInformedBy including in the core in prov-dm


<smiles> 0

<dgarijo> +0

<khalidbelhajjame> 0

<GK> I assume +foires for, - votes against?

<TomDN> +0.5

<GK> +0.5

<stephenc> +1

<SamCoppens> 0

<jun> +0.5

<jcheney> +1 seems useful to cover it for symmetry with derivation even if it is redundant

<dgarijo> what does 0.5 mean? Is that support, partial support or that you don't care?

<zednik> +0

<GK> (dropped off audio briefly there)

<GK> @dgarijo I use +0.5 to mean I lean towards, but happy either way.

pgroth: fairly positive, ok luc?

luc: that's fine.

<TomDN> means that I wouldn't object if it goes the other way

<pgroth> accepted: wasInformedBy included in the core

<dgarijo> @GK: ok, thx!

pgroth: want to confirm other terminology 'core' vs. 'extended', other options?

<pgroth> proposed: Use the term core and extended to categorize the constructs

<TomDN> +1


<khalidbelhajjame> +1

<SamCoppens> +1

<zednik> +1

<smiles> +1

<GK> +1

<tlebo> +1

<dgarijo> +1

<jcheney> +1

<satya> +1

<Luc> does it mean that prov-o starting points becomes prov-o core?

<jun> +1 (for the DM, right? not in every document)

<pgroth> accepted: Use the term core and extended to categorize the constructs

pgroth: these decisions are on prov-dm, what is the impact on prov-o?
... prov-o editors? what do you want to do with this?

tlebo: I'm comfortable changing 'starting point' to 'core', but the time concepts are inconsistent with that

<GK> I propose no immediate impact on PROV-O, unless some one raises an issue. I have no problem with apparent inconsistency, since it doesn't have any technical impact.

jun: we have further categorization for other sections, what about those?

<pgroth> +q

pgroth: It is reasonable for prov-o to keep starting point and mention these are similar to the prov-dm core
... different terminology might help limit confusion, even if there is less parallelism
... prov-dm doesn't have the qualified terms

luc: in section 2.1 of prov-dm, it just talks about the key classes/types/associations, it doesn't get into the notion of time
... that comes in much later, 2.1 is just conceptual, doesn't get into syntax
... it doesn't preclude time being part of the core, shouldn't be a problem with prov-o having that in its starting points

<GK> Luc: "examples of syntax don't come until section 4"

luc: with regard to the qualified relations, there are association classes later

pgroth: include time in the core points?

luc: the representation in dm core doesn't really allow discussion of time, but it isn't really a problem to consider time part of that


pgroth: Keep that terminology? Core for DM and 'starting points' for prov-o?

<khalidbelhajjame> I thought that Luc was suggesting that the core in dm and starting points in provo are the same

<khalidbelhajjame> which I think is the case

<GK> @khalid - I think he said they were not incosistent, just that

pgroth: they aren't the same, since time isn't in the dm core

<Luc> I don't show any attribute in this class diagram

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#component1

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#component2

luc: to clarify: activities in figure do show time, but those attributes aren't really discussed in the core

<GK1> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#term-Activity - st and et are mandatory, so I think their appearance (as distinct terms) in PROV-O start is entirely consistent as it can be due to nature of RDF representation.

pgroth: either we align, or we don't, tim is right, we do know what is part of dm core, maybe we can address alignment later

<jun> +1

<GK> +1

<dgarijo> +1

<khalidbelhajjame> +1

<SamCoppens> +1

<TomDN> +1

Name for Responsibility

pgroth: luc can proceed based on this for now

<tlebo> http://www.w3.org/2011/prov/wiki/Renaming_the_concept_Responsibility

tlebo: name for responsibility is ambiguous
... tlebo too broad, too general, would like to rename to 'delegation'

luc: confirm you are now happy with 'delegation'?

tlebo: choosing between delegation or OnBehalfOf
... I would probably lean toward actedOnBehalfOf, but either that or delegation better than responsibility

<Luc> Responsibility is the fact that an agent is accountable for the actions of a "subordinate" agent, in the context of an activity.

<dgarijo> @Tim: so, if we choose delegation, would we change "actedOnBehalfOf" to something like delegated?

<tlebo> @luc, it is strange.

luc: behalf sounds strange

GK: yes, it does

luc: it isn't really used on its own like that

pgroth: does it have to be one noun?

luc: yes

<tlebo> "Act" ?

<Luc> @tlebo, too close to activity

<tlebo> true

<TomDN> "Action" also too close?

pgroth: delegation does sound good

<JimMcCusker> "actedFor"?

<tlebo> POI: we already have "qualified" v. "unqualifed" naming mismatches, so it's not a show stopper.

luc: delegation has a benefit, and clarifies the presentation in both documents

<tlebo> (wasinformedBy and Communication)

<pgroth> proposed: Use Delegation as the name for what is now Responsibility

<smiles> +1 unless it means changing actedOnBehalfOf to wasDelegatedTo

<tlebo> +1 (it "stands alone" better than "Behalf")

<khalidbelhajjame> +1

<GK> +1

<SamCoppens> +1

<dgarijo> +1

<TomDN> +1


<jcheney> +.999

<jun> +1 (it seems clearer)

<jcheney> +1

<stainPhone> +1

pgroth: simon -- I don't like 'wasDelegatedTo' either

smiles: actedOnBehalfOf is better

<zednik> +1 (Delegation better than Responsibility)

luc: yes, wouldn't want to change it to 'wasDelegatedTo'

<Luc> Delegation is the fact that an agent is accountable for the actions of a "subordinate" agent, in the context of an activity.

luc: can I simply replace 'responsibility' with 'delegation'?

<TomDN> +q

TomDN: keyword subordinate doesn't really work -- you can delegate to a peer

<stainPhone> And what is "accountable" now.

<smiles> Yes, I find it fine just replacing Responsibility with Delegation

<GK> "assignment of responsibility" -> "delegation"

pgroth: simple search replace should be ok

<GK> I think an editorial pass may be needed.

<zednik> from wikipedia, slightly longer definition

<zednik> Delegation (or deputation) is the assignment of authority and responsibility to another person (normally from a manager to a subordinate) to carry out specific activities. However the person who delegated the work remains accountable for the outcome of the delegated work.

<TomDN> I don't disagree, but it is what we define it is

<tlebo> one can look at "senior" with a localized context - even a peer is "senior" if they give a responsibility to a peer (or if a peer "just does it" without the "senior" asking).


<TomDN> so basically it is up to us, no?

<Luc> @zednik: I like your text

<stainPhone> Scribe please? I could not hear what pgroth said.

<pgroth> accepted: use Delegation as the name for what is now Responsibility


<stainPhone> Sorry, lag on irc..

pgroth: who is willing to review when they are ready to go?

<pgroth> prov-dm

<tlebo> @luc, @zednick, I think we should also cover the case where an agent "just does it without asking" for another agent. (altruism)

<dgarijo> I do

<TomDN> I'll review it as well

<khalidbelhajjame> I do


<pgroth> reviewers for prov-o

<dgarijo> so, just to be sure, the document to review is : http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html ??

<SamCoppens> +1

<pgroth> i will

<Luc> @tlebo, ok, i will circulate definition suggestions

<dgarijo> @pgroth: ok, thanks

<SamCoppens> i will

pgroth: Just picking reviewers to review in 1-2 weeks when documents are ready

<Luc> +1

<pgroth> prov-constraints

<smiles> +1

<TomDN> I'll read this one too

<pgroth> prov-n

<SamCoppens> +1

<jcheney> +1

<khalidbelhajjame> +1

<dgarijo> bye!

<Luc> bye

<Luc> @gk

<pgroth> trackbot, end telcon

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/05/24 16:03:59 $

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)

Succeeded: s/gklyne/GK/
Succeeded: s/problem/probably/
Succeeded: s/along/alone/
No ScribeNick specified.  Guessing ScribeNick: Curt
Found Scribe: Curt Tilmes
Default Present: jfuller, TomDN, Luc, +1.518.276.aabb, jun, tlebo, SamCoppens, Curt_Tilmes, dgarijo, khalidbelhajjame, GK1, +44.131.467.aacc, Sandro, +44.789.470.aadd
Present: jfuller TomDN Luc +1.518.276.aabb jun tlebo SamCoppens Curt_Tilmes dgarijo khalidbelhajjame GK1 +44.131.467.aacc Sandro +44.789.470.aadd
Regrets: Paolo Missier
Agenda: http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.05.24
Found Date: 24 May 2012
Guessing minutes URL: http://www.w3.org/2012/05/24-prov-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]