Provenance Working Group Teleconference -- 10 Jan 2013

<trackbot> Date: 10 January 2013

<pgroth> can someone scribe?

<pgroth> weka

<pgroth> mallet

<pgroth> scribe: jcheney

admin

pgroth: WF4Ever meeting so lots of people away

<pgroth> http://www.w3.org/2011/prov/meeting/2013-01-10

pgroth: any objections to minutes?

<pgroth> http://www.w3.org/2011/prov/meeting/2013-01-03

pgroth: minutes from last week

<pgroth> approved: January 3, 2012 minutes

pgroth: open action items:

<Luc> close action-154

<trackbot> Closed ACTION-154 Review the test cases.

pgroth: closing some that were closed last week

<Luc> close action-155

<trackbot> Closed ACTION-155 Review the test cases.

pgroth: stefan working on xml namespace
... paul to send note on implementations, will do today/tomorrow

WG Implementations

pgroth: stephan working on xml namespace
... was hoping for update from stephan (who gets the emails)
... would like to see how to make a report from survey results

<Luc> 10 implementations, 5 vocab extensions

pgroth: Now have 9 impls, 5 vocabulary extensions
... Would like to know what these are

<pgroth> ACTION: send stephan an email to ask for all results of questionaires [recorded in http://www.w3.org/2013/01/10-prov-minutes.html#action01]

<trackbot> Error finding 'send'. You can review and register nicknames at <http://www.w3.org/2011/prov/track/users>.

<pgroth> ACTION: pgroth send stephan an email to ask for all results of questionaires [recorded in http://www.w3.org/2013/01/10-prov-minutes.html#action02]

<trackbot> Created ACTION-158 - Send stephan an email to ask for all results of questionaires [on Paul Groth - due 2013-01-17].

<pgroth> https://www.w3.org/2002/09/wbs/99999/prov-implementation-survey/results

ivan: can see all the responses

pgroth: who has done what? why can't anyone else see it?

<Luc> moj@ecs.soton.ac.uk, dgarijo@delicias.dia.fi.upm.es, dtm@ecs.soton.ac.uk, tdh@ecs.soton.ac.uk, dong.huynh@soton.ac.uk, donght@soton.ac.uk, soiland-reyes@cs.manchester.ac.uk, amir.keshavarz@gmail.com, caron.clement@gmail.com,

<Luc> these are the responders

<Luc> https://www.w3.org/2002/09/wbs/99999/prov-implementation-survey/email-list

ivan: <listing some of the responses>

<Luc> that's my 10 :-)

<pgroth> https://www.w3.org/2002/09/wbs/99999/prov-vocabulary-survey/results

ivan: should be visible to members of "this group" but not sure which group it is.
... vocabulary extensions: 5 for prov-o, none for others

<Luc> irene.celino@gmail.com, dgarijo@delicias.dia.fi.upm.es, Khalid.Belhajjame@cs.man.ac.uk, soiland-reyes@cs.manchester.ac.uk, rpalma@man.poznan.pl,

<Luc> https://www.w3.org/2002/09/wbs/99999/prov-vocabulary-survey/email-list

ivan: what would extension mean for prov-n?

<Luc> you could write an xml schema that extends prov-xml schema (but this is not recommendation track)

pgroth: no results for vocabulary usage

https://www.w3.org/2002/09/wbs/99999/closed

jcheney: the questionnaires are in the "Test Group". Can we move them to our group?

<pgroth> ACTION: pgroth to send ivan an email to put the questionnaires in the right group [recorded in http://www.w3.org/2013/01/10-prov-minutes.html#action03]

<trackbot> Created ACTION-159 - Send ivan an email to put the questionnaires in the right group [on Paul Groth - due 2013-01-17].

ivan: this may be a mistake... will ask sysadmins if it can be fixed
... do you want to see the feature coverage?

pgroth: would like to see feature coverage & interoperability in implementation report

<pgroth> https://www.w3.org/2002/09/wbs/99999/prov-implementation-survey/results

<pgroth> [4:06pm]

@ivan: could you make a screenshot of results and people can look at it off-line?

luc: we have 4 impls that write Entity and 5 that read / write Entity

<pgroth> yeah that's correct

ivan: then averages are meaningful: 4.56 is good

pgroth: for entity, agent we're fine

luc: 6 say no support for invalidation, 1 r/o, 2 r/2. can we assume one reads what the other has written?

pgroth: would be good to see the actual people, so we can check this
... would like to make this public/group readable, and see where there are gaps

ivan: sounds reasonable

luc: just after averages table there are details, but only for two responses - why?

<Dong> that's what I see as well

ivan: can see all 9 rows
... with all responses

pgroth: we need to see what ivan sees asap
... wanted to ask dong what we expect back on constraints
... should they email prov-public-comments or fill out a form or what?

<pgroth> no dong

<Luc> dong?

Dong: Decided to use questionnaire and not email, haven't removed email yet, will do soon
... put link in call for implementations

<pgroth> it's not on the main page

pgroth: will update main page after changes made

<pgroth> i will do that

Dong: need to update front page

Prov-Dictionary

TomDN: prov-dict pushed just before call; everything done except prov-xml
... can be reviewed starting tomorrow
... incorporates results of discussion on mailing list
... can be considered for fpwd after review

<TomDN> https://dvcs.w3.org/hg/prov/raw-file/default/dictionary/prov-dictionary.html

pgroth: please send email / issue tomorrow for review

PROV-AQ

Luc: can we confirm reviewers for prov-dictionary?

pgroth: paolo, stian, james(?), luc, pgroth

<Luc> Paolo, Stian, James (maybe), Luc, and Paul

<pgroth> http://lists.w3.org/Archives/Public/public-prov-wg/2013Jan/0032.html

pgroth: reviewable version is available, questions for review in issue 613
... would like feedback on pingback

<pgroth> Tim, Simon, Luc, Dong and Stian

pgroth: "last call" before prov-aq released as ready for implementation
... deadline for review is thursday next week

Prov-o encoding constraints

<pgroth> http://www.w3.org/2011/prov/track/issues/612

pgroth: comments from two implementors working with prov-o and constraints
... looking for constraints implementable in OWL to be part of prov-o
... this was discussed and resolved not to do this earlier, but this could be an implementation technique
... how to address?

Luc: no consensus for derivation to be transitive; we voted against this and it is not a constraint in the document.
... should review & approve responses, but would be good to tell them this specific point

<Luc> ACTION: Luc to write a response to Kerry why derivation is not transitive [recorded in http://www.w3.org/2013/01/10-prov-minutes.html#action04]

<trackbot> Created ACTION-160 - Write a response to Kerry why derivation is not transitive [on Luc Moreau - due 2013-01-17].

ivan: what is wrong with putting expressible constraints in separate document?
... don't see a case for editing prov-o core document

pgroth: fine if people (in or out of wg) want to encode constraints, but not necessarily part of wg deliverables

ivan: if wg members do this, we can at least publish it somewhere

Luc: need to respond to reviewers, along lines Paul gave
... open questions whether some/all constraints implementable and how
... wg decided to view this as an implementation issue, we can offer to gather experiences with this/axioms suggested by implementors

pgroth: sounds fine, but it seems to come up - should we say this in prov-o or constraints saying this?

Luc: seems reasonable. not sure where. james?

jcheney: could put disclaimer/explanation in constraints, maybe signpost elsewhere

pgroth: could say something in overview, prov-o also

<pgroth> ACTION: pgroth to draft response on owl implementation of prov-constraints [recorded in http://www.w3.org/2013/01/10-prov-minutes.html#action05]

<trackbot> Created ACTION-161 - Draft response on owl implementation of prov-constraints [on Paul Groth - due 2013-01-17].

Test cases response

<pgroth> http://www.w3.org/2011/prov/track/issues/611

pgroth: james responded to questions about constraints; this seems fine as response to that part if wg can endorse
... also asked about prov-constraints test cases
... should they be part of spec?

<pgroth> We

<pgroth> would like to see test suites for the other operational parts of PROV,

<pgroth> in particular for testing inferences separate from validation.

pgroth: in particular, in a normative place.
... and would like further test cases for other documents
... how should we respond

ivan: why do they want it in normative spec?

pgroth: if normative, then better interoperability (they say)

ivan: that's a matter of opinion. otoh, if list of test cases become normative, cannot extend them later, or would have normative & non-normative tests
... if test suite is non-normative, then we have capability to add new tests even when docs published
... had this in rdfa wg
... discrepancies between implementations arose, which were addressed through additional tests

Dong: if we move test cases into normative, are we saying that an impl that passes all test cases are compliant? We would have two definitions
... test cases and original spec
... cannot be sure that test cases cover all constraints.
... would provide false sense of compliance

Luc: other issue is that we don't have formal mappings / equivalence between the serializations
... it could be that some test cases would work in prov-n and not rdf or vice versa.
... not in favor of normative test cases
... do we need other test cases for other specs?

(previous line is pgroth)

pgroth: do we need other test cases for other specs?

Luc: the test suite contains typical examples expressed in prov-n, prov-o. what else could we do beyond having weird examples to exercise syntax?
... more interesting to have useful provenance examples

pgroth: test cases provide example repository, can be used to test compliance with other specs

<Luc> this was a suggestion from the SW coordination group that we have a set of useful provenance examples

pgroth: we seem to have outline of response

Luc: important to try to provide responses promptly because they are trying to implement and may be waiting before submitting reports

<Luc> i have updated page http://www.w3.org/2011/prov/wiki/ResponsesToPublicCommentsCR

pgroth: Luc will do response to Kerry, Paul will respond to general question need one for test cases and their comments on constraints

Luc: created page for responses
... suggest we assign people to address these

pgroth: prov-o (611) essentially same as 612 about encoding constraints in owl, paul will do these
... jcheney will do 611 (constraints)

<Dong> ok

pgroth: 611 (normative test cases) - Dong
... can we do this by monday?

Luc: who will do test cases for other specifications?

pgroth: will do that
... goal do send for approval by wg on monday
... goal to send for approval by wg on monday
... so we can send back on tuesday
... seems uncontroversial so hopefully we can approve over mailing list
... or at least try

<Dong> Monday is fine for my part

<SamCoppens> Bye

<Dong> bye everyone

<pgroth> trackbot, end telcon

- DRAFT -

Provenance Working Group Teleconference

10 Jan 2013

Attendees

Contents