See also: IRC log
<trackbot> Date: 04 October 2012
<pgroth> Scribe: Paolo Missier
<pgroth> http://www.w3.org/2011/prov/meeting/2012-09-27
<pgroth> Minutes of the September 27, 2012 Telecon
<ivan> +1
<CraigTrim> +1
+1
<Dong> +1
<pgroth> accepted: Minutes of the September 27, 2012 Telecon
pgroth: what to do about long-lasting open actions
ivan: it's ok to time out on them and close them, noting that no reply was received
<pgroth> https://dvcs.w3.org/hg/prov/raw-file/dedddeedd578/presentations/wg-overview/overview/index.html
pgroth: finally completed his action, see link above
action 118 (?) also taken care of
<trackbot> Sorry, couldn't find 118. You can review and register nicknames at <http://www.w3.org/2011/prov/track/users>.
action-113 done (issue 446)
action-116 still ongoing
<pgroth> http://www.w3.org/2011/prov/wiki/ProvCRExitCriteria
pgroth: exit criteria were discussed at the latest F2F meeting
<pgroth> http://dvcs.w3.org/hg/prov/raw-file/default/reports/prov-implementations.html
exit criteria are here: http://www.w3.org/2011/prov/wiki/ProvCRExitCriteria
pgroth: these criteria apply to the DM and ontology
ivan: what is an "implementation" in this context
pgroth: using PROV in a dataset,
e.g. markung up a web page
... a vocabulayr for ontologies that extend prov
... SW that generates and consumes PROV models
ivan: fine, suggest adding this phrasing to the wiki page containing the exit criteria
<Dong> @Paul: Sure
ivan: also, when implementations are collected, tag them according to the category where they belong
pgroth: are people happy with
those 3 categories?
... these work for DM and O primarily. What would the criteria
look like for CONSTRAINTS?
... need to demonstrate interoperability.
... proposed criteria: multiple implementations, and show that
they support each of the constraints defined in the doc. This
is done through a catalog of reference test cases that the
implementation must be able to pass
... the implementation must correctly evaluate the test case
against the constraints it is meant to exercise
jcheney: clarification: the criteria include the constraints but exclude the inferences
pgroth: yes, but you probably need to do inferences as well as part of the implementation
<stainNexus7> Who would build those test cases? The wg?
jcheney: are the test cases based on 'validity' which requires inferencing? or is inferencing one possible way to do the implementation
<jcheney> so perhaps the test cases should try to *exercise* the inferences
<Luc> +q
ivan: the constr doc contains inference rules, not just constraints. So are there inferences that will not be tested by the test cases?
jcheney: inferences are a mechanism to define validity, however in the doc we specify that other mechanisms to check validity are fine as well
ivan: that's fine then
Luc: to confirm what jcheney wrote above -- but the test case won't check that inferences have been applied
hook: interoperability should
show that producers and consumers of provenance actually can
use the spec to exchange prov
... the current interpretation of interop does not address
that
pgroth: true for CONSTR, however prov DM and prov O do require demonstration of interop according to the exit criteria
Luc: are we try to gain evidence
for each prov-* individually, or collectively as a whole?
... for example, what does it mean for DM to interop "on its
own"?
pgroth: for DM, you do have to go
through prov-N or prov-O. the impl. should specify which
encoding it supports
... incl. XML
ivan: do we really need an implementation for prov-n which is meant for human consumption? it's not meant to be a machine-exchangeable format
Luc: indeed machine processing initially not the primary goal
<Dong> @Luc, when you mentioned evidence, did you mean that we need to gather proofs beyond submitted answers to the implementation questionnaire?
pgroth: this is good feedback to produce the next version of the exit criteria. we need to be more specific
<ivan> http://www.w3.org/2007/OWL/wiki/CR_Exit_Criteria
<Luc> @dong, submitted answers is what I think we have agreed
<Dong> @luc, thanks, that's good.
<Luc> was the functional syntax document a rec?
ivan: the EC for OWL2 is relevant
because it's got an analog in a functional syntax, which is not
even mentioned in the EC
... because the functional syntax can be mapped to one of the
serializations
<ivan> http://www.w3.org/TR/2009/REC-owl2-syntax-20091027/
ivan: the doc above is analogous to prov-n in our case
<Luc> +q
ivan: so the EC should really apply to prov-o and prov-constr
Luc: should different implementations come from different institutions?
ivan: if EC call for two impl, then yes they should come from different institutions
pgroth: plan to vote on EC next week
<pgroth> ACTION: pgroth to revise exit criteria for next week [recorded in http://www.w3.org/2012/10/04-prov-minutes.html#action01]
<trackbot> Created ACTION-119 - Revise exit criteria for next week [on Paul Groth - due 2012-10-11].
ivan: EC should be fully completed by the time we leave CR, not when we enter CR
<dgarijo> and how do we add an application to the implementation catalog?
pgroth: need volunteers to build test cases for the constraints
<Luc> we can already collect all examples from our specs
<dgarijo> me
<dgarijo> I think Jun was interested as well
<Luc> dong?
<Dong> Sorry, I missed it
Paolo: I can help but can't commit time at this point
<dgarijo> I don't see her here, so I'll contact her to see if she's interested.
<pgroth> me
<Luc> me with Dong
<Dong> Yes
<tlebo> after I see a test case or two, I'll reconsider adding some.
Luc: once they are defined, they should be validated "by expert hand"
<Dong> @Paul, I've understood that it's a part of the work I'm involved in preparing the implementation report
ivan: need a dynamics in place to manage the responses. What is the reporting mechanism?
pgroth: we basically believe them
<Curt> believe and document their assertion
ivan: this means that responses will be managed manually, which may be problematic to scale
Luc: have 100-200 tests at the moment, we should have a simple mechanism with an ID per test...
ivan: a basic mechanism should be defined, we must specify how implementors are expected to report back
pgroth: going through the
list....
... 520 left till next time as there was discussion
... (isolating the issues that received feedback and
discussion)
<pgroth> ISSUE-531, ISSUE-528, ISSUE-517, ISSUE-501, ISSUE-516, ISSUE-514, ISSUE-513, ISSUE-511, ISSUE-510, ISSUE-512, ISSUE-497, ISSUE-515,
pgroth: the issues above have reached resolution
<pgroth> proposed: ISSUE-531, ISSUE-528, ISSUE-517, ISSUE-501, ISSUE-516, ISSUE-514, ISSUE-513, ISSUE-511, ISSUE-510, ISSUE-512, ISSUE-497, ISSUE-515 are confirmed to be resolved
<ivan> +1
<tlebo> +1
<dgarijo> +1
<jcheney> +1
<Dong> +1
+1
<pgroth> accepted: ISSUE-531, ISSUE-528, ISSUE-517, ISSUE-501, ISSUE-516, ISSUE-514, ISSUE-513, ISSUE-511, ISSUE-510, ISSUE-512, ISSUE-497, ISSUE-515 are confirmed to be resolved
<ivan> issue-509?
<trackbot> ISSUE-509 -- Data Model Figure 5 -- open
<trackbot> http://www.w3.org/2011/prov/track/issues/509
pgroth: see issue-509
Luc: earlier versions of the docs show capitalized classes, and relationships not capitalized. that led to inconsistencies
Luc
Luc: we addressed by cap "class
level" elements and nocap for "instance level" elements
... diff. notations use different styles. in prov-n nothing is
cap
... so we will be inconsistent anyway whatever change we
make
... in prov-o class derivations are cap, for instance. There is
no solution that works for all of them
... we tried to make prov-dm consistent with itself
ivan: as the one reopening the
issue: looking at the primer on its own. because of its role,
felt that consistency was important
... the primer has dual syntax for examples. in prov-n there is
no cap, while turtle is also consistent with prov-o
... but fig. after sec 2 uses an inconsistent cap mode, and
that is not explained. so proposed to make it consistent with
prov-o
... so just asking to make the figure consistent with one
syntax in the text.
Luc: the figure uses the prov-dm
convention
... it's a class diagram, not an instance.
ivan: the figures in the primer
may differ from those in the DM. because it's the primer,
readers won't appreciate the alignment with prov,
... rather they will be confused by the change in cap style
<dgarijo> +1 to what paul suggested.
pgroth: we should be using the diagram in prov-o instead, it's not UML but it's "classes and properties" and may work better here
ivan: happy with that
<dgarijo> http://dvcs.w3.org/hg/prov/raw-file/tip/ontology/diagrams/starting-points.svg
<tlebo> sounds good
<pgroth> accepted: use a modified version of the prov-o starting points figure in the primer
<tlebo> bye! Thanks, Paul.
<Dong> thanks, bye all
<khalidBelhajjame> bye
<pgroth> trackbot, end telcon
This is scribe.perl Revision: 1.137 of Date: 2012/09/20 20:19:01 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) No ScribeNick specified. Guessing ScribeNick: Paolo Found Scribe: Paolo Missier Default Present: [IPcaller], ivan, Paolo, Luc, Curt_Tilmes, CraigTrim, tlebo, Dong, stain, jcheney, +1.818.393.aabb, dgarijo, +1.818.731.aacc, +1.818.731.aadd, +1.818.393.aaee Present: [IPcaller] ivan Paolo Luc Curt_Tilmes CraigTrim tlebo Dong stain jcheney +1.818.393.aabb dgarijo +1.818.731.aacc +1.818.731.aadd +1.818.393.aaee Regrets: Tom De Nies Agenda: http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.10.03 Found Date: 04 Oct 2012 Guessing minutes URL: http://www.w3.org/2012/10/04-prov-minutes.html People with action items: pgroth WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]
