W3C

- DRAFT -

Provenance Working Group Teleconference

15 Mar 2012

Agenda

See also: IRC log

Attendees

Present
Luc, Paolo, MacTed, Curt_Tilmes, dgarijo?, +1.315.723.aaaa, tlebo, +49.302.093.aabb, +1.443.708.aacc, jun, jcheney, SamCoppens, GK, +49.302.093.aaee, olaf, khalidbelhajjame, [IPcaller], kai
Regrets
Paul, Groth
Chair
Luc Moreau
Scribe
smiles

Contents


<trackbot> Date: 15 March 2012

<Luc> http://www.w3.org/2011/prov/wiki/ProvDMWorkingDraft5

I can scribe

<dgarijo> I h ave to leave very early today, sorry :(

<scribe> Scribe: smiles

<dgarijo> @smiles: thanks

admin

Luc: any other business?

dgarijo: Pending reviews on ontology, please send

<Luc> http://www.w3.org/2011/prov/meeting/2012-03-08

<dgarijo> +1

Luc: Express support for minutes

<Curt> +1

<Paolo> +1

<tlebo> +1

+1

<Luc> ACCEPTED: the minutes of 2012-03-08 teleconference

<Luc> http://www.w3.org/2011/prov/track/actions/open

Luc: Action on Paul to produce revised version of PAQ, sends his regrets for today

Graham?

<GK> I'm not dialled in yet. No progress on action AFAIK.

Action to Tim to ensure WD4 comments addressed

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

Tim: Yes, completed

Luc: Reminder, next week same as this, tc is one hour earlier
... Also need scribes for future tcs

f2F3

Luc: Good support in poll, so will proceed with 22-23 June in Santa Barbara
... preceded by IPAW, James Frew will provide hotel suggestions soon

PROV-O

<dgarijo> http://www.w3.org/2011/prov/track/products/3

<GK> (Unfortunately, the date clashes with an important family commitment)

<tlebo> http://aquarius.tw.rpi.edu/prov-wg/prov-o

tlebo: PROV-O completed iterating on HTML document
... now automatically generating cross-reference reflecting Jun/Khalid's proposed structure
... request any feedback from the WG, please raise issues directly using tracker
... Now preparing to assign people to write narrative for individual sections
... Daniel done great reviewing of sections, please help bring issues to closure where replied to

Luc: Been tremendous work on ontology and document

<GK> It looks v good to me. Well done!

Luc: Is there a stable ontology?

tlebo: Needs to be in sync with DM WD4, so may be couple of changes to ontology (and automatically HTML)
... currently using PROVRDF to track mapping to DM

<dgarijo> @All: please check out http://www.w3.org/2011/prov/track/products/3. There are 8 or 9 issues pending review, so if you have raised them please see if they can be closed.

Luc: HTML also looks great

<GK> Does anyone think we can get it to REC with the title unchanged?

<Luc> ;-)

<tlebo> @GK ;-)

dgarijo: Link pasted above issues against ontology, so if you raised issues please check whether they can be closed

Luc: Suggest for next telecon, have summary of issues open and who raised them

dgarijo: Have already sent individual emails on issues

<jun> @tlebo, is there a link to check all issues related to provo.html?

Luc: Also mail list to add pressure

<tlebo> @jun, yes. http://www.w3.org/2011/prov/track/products/10

<jun> @tlebo, thx! :)

<tlebo> "PROV-O HTML" in pull down on tracker.

dgarijo: Have also divided issues to those on ontology, those on HTML, those on mapping

prov-primer

<Paolo> oops that may have fallen through, sorry about that

Luc: aim to have drafts on all docs for end of month, including primer

prov-dm

<Luc> http://lists.w3.org/Archives/Public/public-prov-wg/2012Mar/0230.html

smiles: Will prepare for end of the month, with input from Paolo on inconsistencies with DM, from Stian on inconsistencies with PROVO

<tlebo> +1 @macted, we should stick to eggs and unchewed gum.

MacTed: express examples in terms of real world things, as expiry and change over time unclear with current examples

Luc: three proposals circulated, second and third had support
... first did not have consensus, cannot be adopted, so will try to change and discuss
... can incorporate second, third proposal in editors draft, can raise issues against in tracker in usual way

GK: First proposal needs motivating example

<Zakim> GK, you wanted to say I think a concrete example for wanting prop 1 would help

<dgarijo> +q

<GK> I'm hearing, but see chat if you can't hear me

<Luc> we can't hear you

GK: <some random sounds>

<GK> Yes

<GK> Well, that was it -- a concrete example that shows...

<GK> ... why it's useful to have entioty expiration.

Luc: Are examples in proposal not satisfactory?

<GK> I didn't find them compelling ... need to rechack.

<Luc> ACCEPTED: to rename wasStartedBy(activity,activity) in wasStartedByActivity(activity,activity)

<Luc> AcceptED: to formulate start and end of activities independently of responsibility and agents. Hence, start and end of activity would no longer be a form of activity association. Instead of an agent, an optional entity trigger would be allowed.

<dgarijo> I had a question about proposal 2 :(

<Paolo> http://www.w3.org/2011/prov/wiki/ProvDMWorkingDraft5

<dgarijo> well, the subject has changed..

<dgarijo> -q

Paolo: Work in progress on WD5, some of to do list are being tracked

<Paolo> http://dvcs.w3.org/hg/prov/raw-file/default/model/working-copy/wd5-prov-dm-components.html

Paolo: As seen on agenda, component structure now excludes core/component divide

<Paolo> http://dvcs.w3.org/hg/prov/raw-file/default/model/working-copy/wd5-prov-dm-quote.html

<scribe> ... New structure reflects discussion on list, please look at it and raise objections

<GK> To follow up my previous comment, the examples given explain what might trigger an expiration assertion, but do not explain why this is useful or needed for provenance (i.e. assertions about things that have happened - if the provenance is asserted, then presumably referenced entities existed at that time - if they disappeared later, is that really something we need to know?)

UNKNOWN_SPEAKER: Quotation / orginal source controversy - could not see where distinction is

<Paolo> http://dvcs.w3.org/hg/prov/raw-file/default/model/working-copy/wd5-prov-dm-components.html#term-value

<dgarijo> I really have to leave. Apologies...

<dgarijo> bye!

<kai> @dgarijo: i take over :-)

UNKNOWN_SPEAKER: Literals are now expressed as syntactic construct, all we say about values is that they have a type: qualified name, XSD type

<Paolo> http://dvcs.w3.org/hg/prov/raw-file/default/model/working-copy/wd5-prov-dm-collections.html

UNKNOWN_SPEAKER: Collections is work in progress, tried to accommodate open issues and agreed with Stian what to include, final iterations with proposed version in above link
... primitive relations regarding insertion, deletion; data types for collections; additional relations to simplify expressions with multiple key-value pairs

<Paolo> http://dvcs.w3.org/hg/prov/raw-file/default/model/working-copy/wd5-prov-dm-plan.html

UNKNOWN_SPEAKER: Agents in association: proposed new formulation for wasAssociatedWith; plan to open for discussion; if no controversy will be in next draft

Luc: For all edits, discussed issues with people, will now include in WD and notify WG when done
... still to address feedback on WD4 to incorporate in document, go through remaining issues; revision by end of the month

smiles: What is controversial, under development?

Luc: Once collections issues addressed, only presentational issues remain, so stable model
... Proposal 1 above is still controversial: expiry

Paolo: Agreed, everything before expiry proposal, everything settling down
... Not everything in collections draft is uncontroversial but getting there

<khalidbelhajjame> @smiles, do you intend to speak about collections in the primer?

<scribe> ... New pieces in BNF for optional arguments, impact on PROVN examples

Paolo: should be pedagogical text on collections in primer

Luc: Finished agenda, but useful to discuss expiry

<GK> I can hear OK, but not speak

<GK> To follow up my previous comment, the examples given explain what might trigger an expiration assertion, but do not explain why this is useful or needed for provenance (i.e. assertions about things that have happened - if the provenance is asserted, then presumably referenced entities existed at that time - if they disappeared later, is that really something we need to know?)

<khalidbelhajjame> Is the objection against the example or the notion of expiration?

MacTed: An entity may be referred to after being destroyed
... what do we mean here?

<tlebo> unbroken eggs and unchewed gum!

<Curt> "invalid" = no longer can be 'used' by an activity

<GK> The notion. In my case, I'm not seeing why it's useful for provenance. Use-case?

<GK> I'm not opposing it, just not supporting it.

<GK> Why do we care?

MacTed: Table example (logs to table, table broken up and burnt); in contrast, resolution on web gets latest revision, so not really identifying one entity

<tlebo> Let the web people deal with the web problems, not the provenance people. "not able to be used again" is a useful notion (eggs and gum)

<MacTed> I don't know who [IPcaller] is ... but hopefully they can unmute themselves when they figure it out

Paolo: Want to demarcate lifetime of entity, existence to non-existence (becomes another entity), but can refer to forever

<Curt> The examples *don't* discuss things that are 'destroyed', just things that are no longer accessible/available.

<tlebo> having the symmetric to "Generated" seems to be rather nice.

<GK> So what does the entity URI identify when the entity ceases to exist?

Paolo: Nothing to do with real thing getting destroyed

jcheney: Important to distinguish entity which is information valid during a time interval from thing described by that information which can be created, destroyed
... start and end of both are valid to talk about

<khalidbelhajjame> The entity is invalidated not when the thing is distroyed, but when it the entity is no longer reflect the thing it represents.

MacTed: An entity is a thing which is described, not the description

<Paolo> +

Luc: Description holds or not

<GK> If there's any meaning here, I think its to do with validity of statements *about* the entity.

jcheney: Lack of entity expiry in semantics not difficult in itself, but awkward if want to say "table stopped being in room on Wednesday"

<Curt> If the characterization of an entity includes the attribute that it be available through a URL, then when that URL no longer refers to that content, then entity is 'destroyed'.

MacTed: Entity is the thing not the desciption of the thing

<GK> I suppose it might be seen as implying a contextual constraint on statements like "entity dc:creator Foo"?

Luc: Agree with GK that it is about validity of statements about the entity

<Curt> an entity is a set of attributes that characterize something When one of those attributes changes (any of them), you are talking about a different entity.

<Paolo> I am going by this definition: http://www.w3.org/2011/prov/wiki/FormalSemanticsStrawman#Entities

<tlebo> @macted, yes.

<GK> If we think of provenance in terms of entities interacting with processes, then the points at which provenance applies are exactly when those interactions occur?

Paolo: Should follow definitions in strawman formal semantics
... If no compelling reason to have concept, then more trouble than its worth

<GK> ... or in spans delimited by those interactions?

<jcheney> Current version is at http://www.w3.org/2011/prov/wiki/FormalSemanticsWD3#Entities

MacTed: An entity has internal integrity solidified by attributes which are immutable, depending on timeslice
... e.g. a table is an entity which is a timeslice of the wood comprising it

<GK> I *would* see some point if there were an action that "expired" an entity.

<jcheney> @MacTed, please raise issues against semantics if what is there is unclear or could be improved

Luc: Do not want unbounded discussion, so need prepration, proposals

<GK> The problem is I don't oppose this idea, just don't support it :)

Luc: Can MacTed, others formulate notion of end of lifetime in your terms?

<khalidbelhajjame> -q

<GK> I think the onus is on those who think the concept is useful to come up with compelling examples.

<GK> They didn't compel me.

<khalidbelhajjame> +q

<Curt> I agree - they are not compelling

<GK> ... the examples given explain what might trigger an expiration assertion, but do not explain why this is useful or needed for provenance (i.e. assertions about things that have happened - if the provenance is asserted, then presumably referenced entities existed at that time - if they disappeared later, is that really something we need to know?)

khalid: Do we need notion of destroy/expire in any application?

<MacTed> if a Great Master's painting (original canvas) is burned in a museum fire, this is important knowledge when it apparently surfaces later at auction...

<Curt> I've got to go, but I'll think about this some more...

<GK> I see here there's a weak consensus for leaving expiration as proposed. Nobody opposed it, just lots of people didn't care.

<MacTed> (concrete, real world, not rotten eggs...)

Paolo: Will come up with a couple more examples, seems odd to express only one end of timeslice

<GK> @paolo I mentioned earlier that if there were a specific activity-related event that causes an entity toi be invalidated, that would make sense.

Luc: MacTed please share definition by email

<GK> @Luc - burned poainting - se eprevious comment :)

Luc: GK, MacTed's example is interesting one to motivate

<GK> Bye.

<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

trackbot, end telecon

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/03/15 16:03:58 $

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: smiles
Inferring ScribeNick: smiles
Default Present: Luc, Paolo, MacTed, Curt_Tilmes, dgarijo?, +1.315.723.aaaa, tlebo, +49.302.093.aabb, +1.443.708.aacc, jun, jcheney, SamCoppens, GK, +49.302.093.aaee, olaf, khalidbelhajjame, [IPcaller], kai
Present: Luc Paolo MacTed Curt_Tilmes dgarijo? +1.315.723.aaaa tlebo +49.302.093.aabb +1.443.708.aacc jun jcheney SamCoppens GK +49.302.093.aaee olaf khalidbelhajjame [IPcaller] kai
Regrets: Paul Groth
Agenda: http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.03.15
Found Date: 15 Mar 2012
Guessing minutes URL: http://www.w3.org/2012/03/15-prov-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]