Provenance Incubator Group Teleconference -- 24 Sep 2010

<trackbot> Date: 24 September 2010

I can scribe.

But please do all the technical things for me.

<smiles> I'm afraid I will have to leave the call early: 4.30pm

<YolandaGil> Scribe: Christine

Yolanda; Will discuss putting together the final report and the defn of provenance.

<pgroth> did you see that mendeley came out with public sharing

Yolanda: Paolo has been working on importing Mendelay collection to B (?)

<Paolo> yes

<YolandaGil> http://www.w3.org/2005/Incubator/prov/wiki/Recommendations_for_scenarios

Paolo: Need a protocol to keep updated.

<Paolo> @Paul does that mean the 10 mmbers limit is gone?

<Paolo> @Paul will read that again later

Simon: Looked a scenario for core recommendations different to news aggregator - 2 different aspects - a procedure that has been enacted so can compare with requirements - perhaps need a standard whay to represent a procedure that has been enacted.
... the second one - commonlaity of derivation of two sources - do 2 things come from the same source?
... Some aspects re management of provenance might go beyond what we want to recommend - but some things like selective disclosure
... also guaranteed usability of provenance

? Some comments to provided by email. 2nd scenario provenance of provenance but 3rd scenario might lead to a recommendation - usable as a record

Yolanda: Provenance as a record - not using that exact terminology - need to be able to express the provenance of provenance - maybe refine this

Is selective disclosure a recommendation?

Yolanda: propose as a separate recommendation
... See recommendation 3 - I thought we agreed - 1st scenario core requirements, 2nd adds to that and 3rd adds to that - assume if in 1st scenario will be relevant to the others

Discussing mechanism to show provenance has not changed

Immutability vs, non-repudiation of provenance

Signable provenance for non-repudiation

Conclusion? - the 3 management points are already covered by recommendations 2 and 3?

Yolanda: yes, 2, 3 ,4

[inaudible]

<pgroth> +q

<pgroth> ack

Yolanda: Re disease outbreak scenario - those 2 recommendations were the major ones
... We did not discuss last week - the use of policies mentioned in the scenario
... This scenario highlights more data aggregation (from different sources) - should we have an item that specifically states this - provenance may be more detailed - more vocabulary may be needed?

Irini: For disease outbreak scenario - important to mention data integration - database area there has been a lot of work re provenance at different levels of granularity - queries - we should look at the work done for relational databases

Yolanda: We have not highlighted time or the issue of handling different versions of information as it changes over time - wonder if it should be captured somewhere

? question about time - time of artefact may not match the time you record the provenance

Yolanda: in the 2nd recommendation we mention time
... we need to be more specific which time

Paolo: Need to have the distinction to work in different ways

? May be captured by recommendation 4

Yolanda: We will make that more explicit

Paul: related to time is the notion of versioning
... Does it come up in the business case scenario?

Yolanda: yes - in the last paragraph of the disease outbreak scenario

Paul: recommendation 5 is very specific but 1-3 recommendations (coarse grained provenance) then there is versioning

Yolanda: Maybe useful to have something about versioning

Paul: or mention versioning in the explanation of recommendation 5

? in the OPM profile space - would argue should go out to standard ontologies

re versioning

? business contract scenario raise versioning

<ssahoo2> Is there is a connection between granurality of provenance and versioning?

Paul: Need to resolve what we are going to say re versioning

<smiles> sorry, but i'd better go. thanks, simon

<ssahoo2> I think we should mention versioning explicitly in the recommendation

Yolanda: If look at recommendations as they are now - we need a theme and all these would be themes - maybe provenance is important for the web and recommend pursue provenance in these ways ...
... Different priorities for recommendations?

<Paolo> +1 for attaching priorities

<pgroth> +1

Yolanda: Core of provenance representations for the web? and maybe extensions

<ssahoo2> I think recommendation 1,2,3, and 7 have higher priority than others

Satya: how to represent the provenance information (1) and (2) how to access provenance and (7)

Sam: Like the idea of versioning - might be good to keep manageable by prioritising them

<ssahoo2> http://www.w3.org/2005/Incubator/prov/wiki/Recommendations_for_scenarios

<YolandaGil> http://www.w3.org/2005/Incubator/prov/wiki/Recommendations_for_scenarios

Paul: Priority 1 is rec 3 then 1 then 2

Paulo: we need to revisit some of the foundations

Yolanda: perhaps an email to the list with details

<Paolo> I just got temp disconnected

<Paolo> but have been on the phone

Luc: broadly in agreement with Paul re priorities - but need to do at a higher level - provenance is very broad - solve core but not everything

Yolanda: which items would be more amenable to be part of the core?

Luc: agreed with Paul's prioritization

<SamCoppens> I would suggest 1, 3, 2 for the core

Jun: rec 4 - 7 seem to be quite specific - before we can do that we don't have the foundation for the semantic web community - standardisation of representation of provenance info - 3 and 1 are my priorities

<pgroth> 3 isn't query, it's access

Paolo: rec 1, 2, 3 are the baseline/foundation - placeholders 1 for data model, 2 .. 3 being open to other models - they are the baseline

<pgroth> would hate to specify a query mechanism

<Jose_> IMO 1,3,2 first, then 5,6,4,7

<JimM> core, integration with DC, versions, metadata, records, access control

Jim: generally agree that there is some core of things going through processes - integration with external metadata needs to be there - records is something we have not addressed a lot and is a key thing (particularly for business)
... some things around reasoning and granularity

<JimM> what's out for standardization - reasoning between witnesses, assessing reproducable research, inference over granularity

Irini: agree with Luc and Paolo - priorities are different for different contexts - e.g. 5 then 3 for linked data SPARQL

<pgroth> +q

Yolanda: clear focus of the group 1, 2 ,3 in different orders - Paolo summarised quite well - data model, query model and open model

<Luc> +q

Yolanda: other areas require more work, more research

<Luc> is it query model or query language?

<Luc> designing a query language is a hard task!

Paul: a query model/approach - I though the recommendation was about access or making information available - prefer not to call query
... difference between access and query language

Luc: agree

<ssahoo2> access vs. retrieval mechanism?

Paolo: access is too broad - there should be an element of selectivity - to be able to specify what you want - query too specific

Irini: explaining previous comment ... sparql language can be used for queries to produce provenance information - want fined grained so I can reproduce my results...

Yolanda: pls write a paragraph to illustrate that different recommendations may be relevant for different circumstances

<YolandaGil> http://www.w3.org/2005/Incubator/prov/wiki/What_Is_Provenance

Yolanda: useful call - discuss provenance defn next time - we have a page on the website (see agenda) - it has a few definitions - question is whether we will be able to come up with one defn as a group

Christine: for outside the provenance community it would be helpful to have one definition

Yolanda: we will discuss this next week - also look at what provenance is not
... thank you

<YolandaGil> Christine: i'll take on from here, many many thanks for scribing!!

<YolandaGil> ScribeNick: Christine

<YolandaGil> trackbot, end telcon

- DRAFT -

Provenance Incubator Group Teleconference

24 Sep 2010

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output