Provenance Working Group Teleconference -- 26 Jul 2012

<trackbot> Date: 26 July 2012

<Luc> Scribe: Curt Tilmes

<Luc> @sandro, zakim does not seem to see anybody on the phone

<Luc> sandro, are you present?

<Curt> Agenda: http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.07.26

<Curt> Scribe: Curt

<Luc> proposed: to approve minutes of last week's call

<smiles> +1

<TomDN> +1

<zednik> +1

<khalidBelhajjame> +1

+0 (not present)

<Paolo> +1

<jun> 0 (not here)

<hook> +1

<tlebo> +1

<jcheney> +1

<Luc> accepted: to approve minutes of last week's call

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

Luc: two actions on paulo, 1 on paul

<sandro> (previous meeting running late, sorry.)

<Luc> @sandro, zakim does not seem to see anybody on the phone

PROV-DM/O/N LCWD

Luc: documents were released as last call working draft
... several announcements have been made about them

<pgroth> http://www.w3.org/2011/prov/wiki/OutreachInformation

pgroth: emails have gone out, blog post about the release
... may want to refer to a specific blog post depending on outreach audience and focus
... Who will reach out and to where?

<Paolo> BTW I sent the announcement to the DataONE project "community" list

<jcheney> dagstuhl list(s)?

pgroth: Sent to several lists, Curt posted to several, others?

<jun> Anyone to pub-lifsci list?

<jcheney> ok

Luc: Yes, James, send to dagstuhl
... Jun, yes, please send there

<pgroth> provenance-challenge?

jun: I will

<pgroth> yeah

prov-constraints

Luc: constraints released internally, check on reviews
... Simon has submitted a review

<tlebo> me, tomorrow.

stain: I will review, but I've just started reading it
... May not finish by friday
... probably need another week

pgroth: I will review, probably complete by tomorrow

<pgroth> https://github.com/pgroth/prov-constraints-validator-spin

pgroth: I will try to implement it
... questions on interactions between prov-n and constraints
... talk about those, or include in review?

<pgroth> ok

<TomDN> I'll go over it again tomorrow as well, but don't have much time left in my schedule.

Luc: Include those in the review

<pgroth> it is heavy going

PROV-O PrimarySource

Luc: Thanks to reviewers -- it is not a straightforward document

<jcheney> We can do a lot to make it easier to understand - it was not straightforward to write either :)

<jcheney> (at least my parts)

tlebo: Comments on latest draft about naming inconsistency between prov-n and prov-o

<stain> @pgroth is that SPIN you mentioned? Matt Gamble suggested to use that to me for this purpose just 10 minutes ago ;)

<pgroth> @stain check out the github

tlebo: prov-n allowed to type the relation, prov-o uses a class

<pgroth> @stain it's working well

tlebo: proposal to rename class prov:Source to prov:PrimarySource prov:qualifiedSource to prov:qualifiedPrimarySource
... Rename will fix inconsistency

zednik: There might be some confusion, need to clarify that primary source isn't an entity, it is a relationship between entities

pgroth: This isn't really a constraint, just an issue for validation / validator

<pgroth> ok good

Luc: a derivation can't be an entity by definition

<Luc> Proposal: Rename prov:Source to prov:PrimarySource, prov:qualifiedSource to prov:qualifiedPrimarySource

<tlebo> I'm glad that prov-c clarifies the confusion on Entity vs. PrimarySource (and Quotation and Revision)

<stain> @pgroth so https://github.com/pgroth/prov-constraints-validator-spin - looks good

<tlebo> +1

<zednik> +1

<stain> +1

<TomDN> +1

<jcheney> +1

<hook> +1

<smiles> +1

<jun> +1

<Paolo> +1

<sandro> +1

<Luc> Accepted: Rename prov:Source to prov:PrimarySource, prov:qualifiedSource to prov:qualifiedPrimarySource

<pgroth> yes

Luc: Is such a rename part of the technical content of the spec?

<stain> sounds like an issue.. imagine <body> changing name to <content> after last call of HTML

sandro: Would anyone happy with the spec become unhappy because of this change? This trivial change is probably ok.

<stain> (or more like <blockquote> to <quote> !)

sandro: Might want to take a closer look

<stain> we can pay a third party to complain about it ;)

<pgroth> +q

Luc: It might introduce a large delay to re-do the last call, how should we handle this?

sandro: I would think this is minor enough not to need a new last call, but we might need to look at it

<stain> @Sandro +1, don't do a new LC because of this!

pgroth: We are proposing this change to respond to comments

<tlebo> +1 @pgroth, we are responding to their comments.

sandro: The issue is conformance, test cases, etc. You are allowed to make editorial changes and fix bugs, but not change design decisions.

<Luc> it is a bug fix, aligning names

<tlebo> this is very bug-like.

Luc: We are just trying to fix the bug -- an inconsistency between the documents.

sandro: It seems crazy to delay things by 4 weeks for such a minor thing.
... This seems like a reasonable exception

Luc: editors should implement this change and note it as a bug fix
... We should keep good records about the approach
... Ok, Tim?

<tlebo> PROV-O latest draft has http://aquarius.tw.rpi.edu/prov-wg/prov-o#changes-since-wd-prov-o-20120724

tlebo: yes

Working Draft XML

<Paolo> horrible home networking problems, can't hear a thing -- going to check out

zednik: Per email, there is an XML schema from Luc. Starting by reviewing that

<Luc> http://lists.w3.org/Archives/Public/public-prov-wg/2012Jul/0362.html

zednik: Reviewing terms, creating examples similar to prov-o, mapping things to XML
... providing feedback on XML schema's ability to support the examples
... A google spreadsheet is organizing activities and we will provide feedback on the terms

Luc: Are the terms assigned yet?

zednik: Not all are assigned yet, will get them assigned by the end of this week
... Anyone that wants access to the spreadsheet, select "Request Access" and I'll grant it

Luc: Timeline?

zednik: Need to consult group, haven't nailed down complete schedule yet, but some examples are underway

Luc: We agreed on a deadline to have the notes ready for last call review internally by ~mid-November

<Luc> https://docs.google.com/spreadsheet/ccc?key=0An15kLxkaMA3dEtpbTNlNEQ4eGJqQUtKVXFVekFxR1E#gid=0

Luc: To do that, by a month earlier, we need to have a completed schema, agree on it by September
... Need examples by end of August

zednik: Try to get examples produced by mid-august so we can discuss and revise if needed

Luc: Paul/Luc will be on holiday second half of august

sandro: Someone else can chair if neither co-chair is available

Luc: In the next week or two we should agree on the time line. While paul/luc are away, the regular telecon could be used to discuss the XML and examples

Preparing the Call for Implementation

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

<pgroth> or know about

Luc: Inviting members of the group to add implementations you are aware of to that wiki page

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

Luc: paul and I are drafting the charter extension request -- need that page as evidence of uptake of the spec
... We need to identify a set of features that must be implemented by two implementations
... We need to start thinking about the exit criteria now, so we are ready when we want to move to candidate rec.
... How should we address this?

<pgroth> +q

zednik: To clarify -- we need to identify a set of features that must be implmemented. Is that of all features, or some minimal set?

Luc: We will look at each feature and want to list which implementations 'support' that feature.

<sandro> (But it could be different implementations for each feature)

Luc: We want to get at least two implementations for each feature
... We also want to get a pair, including a producer and a consumer for each feature

<stain> but what is a 'feature'? How granular?

pgroth: We should enumerate features based on section -- we should put up a wiki page to gather the list

zednik: We had the same issue of enumerating features for the XML review -- that list may be incomplete, but may be useful for this
... Others are invited to review that list and see if any are missing

Luc: That is a reasonable approach -- we should also list features that are document specific
... Each CR will have its own list of features (many of course will be the same)

<pgroth> sure

<pgroth> the rules

Luc: For constraints, e.g. each constraint will become a listed feature

pgroth: Agreed, that was just a for starting point
... We need to support the full vocabulary.

<pgroth> use the exit criteria page?

<pgroth> ok

Luc: Paul, will you create the starting point page and we'll discuss on email in the coming weeks?

<jcheney> The RDF semantics is a recommendation - what were its exit criteria?

<sandro> jcheney, all of RDF was done together, not separately.

zednik: I've started a form in google docs for a questionairre on the implementation/support/etc.
... I'll share the form and request feedback by today or tomorrow

<pgroth> that's fine for constraints

sandro: Other groups have done this with executable test cases, each test case is a feature, machine readable results were used to determine coverage

<pgroth> maybe

<zednik> http://www.w3.org/2006/07/SWD/SKOS/reference/20090315/implementation.html

zednik: SKOS is the model we are trying to follow

<sandro> thanks

<dgarijo> I think this approach is cool!

Luc: Constraints may be handled differently -- we might want graphs that violate or are compliant with the spec

- DRAFT -

Provenance Working Group Teleconference

26 Jul 2012

Attendees

Contents

PROV-DM/O/N LCWD

prov-constraints

PROV-O PrimarySource

Working Draft XML

Preparing the Call for Implementation

Summary of Action Items

Scribe.perl diagnostic output