F2F Meeting, Stata Center -- 09 Nov 2012

<pgroth> Scribe: Curt Tilmes

Admin

<pgroth> http://www.w3.org/2011/prov/meeting/2012-11-01

<tlebo> =1

<Curt> 0 (not present)

<tlebo> +1

<hook> +1

<ivan> (was not present)

<tlebo> (that actually works, since I was the first vote...)

<pgroth> approved minutes of the 01 November 2012 telco

Where we are at

<Luc> approved: minutes of the 01 November 2012 telco

<Curt> pgroth: we are in great shape!

<Curt> pgroth: will discuss documents on rec. track

<Curt> pgroth: most issues closed or will be momentarily

<Curt> pgroth: need to follow w3c process and do due diligence

<Curt> pgroth: document everything clearly

<Curt> pgroth: CR period will focus on implementations

<Curt> pgroth: both finding other folks to implement as well as working on implementations ourselves

<Curt> pgroth: we must show that we implement these specs

<Curt> pgroth: need coverage of all the features

<Curt> pgroth: reach out to people, engage others, push notes out, FAQ, etc. for outreach to implementers

<Curt> pgroth: it has been a long hard slog to get here, need to keep up momentum and let people know what we've done

<Curt> Luc: we've done amazing work since the last meeting there has been serious progress, now we need to finish

<Curt> Luc: need to promote the work that has been done

<Curt> GK: getting specs out is the start, we now hope the wider community will pick things up

<Paolo> sorry to be a pest: I think the phone mic goes to sleep even with short pauses so now it's all very on/off -- hard to follow. only continuous voices come across clean

<Curt> hook: this is a time to focus on implementations -- two serializations (PROV-O, PROV-XML) are each distinct encodings, distinct implementations

<Curt> hook: current definition is loose

<Curt> pgroth: in terms of implementation, we are looking for usage. A markup of a web site is an implementation

<Curt> pgroth: we are also looking for things that generate, consume, validate constraints, etc.

<Curt> pgroth: we will see people use PROV as the basis for other work

<Curt> pgroth: our exit criteria count data marked up, vocab. extensions, applications each as implementations

<Curt> GK: do extensions help us with CR exit criteria?

<Curt> pgroth: yes! similar to SKOS, we want to verify that people are using the work. That includes markup and extensions

<Curt> Luc: obviously, we are looking for applications to generate and consume provenance -- those really demonstrate interoperability

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

Candidate Recommendation for prov-dm

<Curt> pgroth: status outstanding issues

<Luc> http://www.w3.org/2011/prov/track/products/2

<ivan> issue-482?

<trackbot> ISSUE-482 -- [external question] bundle IDs on insertion, context -- pending review

<trackbot> http://www.w3.org/2011/prov/track/issues/482

<Curt> pgroth: haven't received acknowledgement from externtal reviewer satra

<Luc> http://lists.w3.org/Archives/Public/public-prov-comments/2012Oct/0019.html

<Curt> Luc: He has acked.

<GK> Re my unease - the exit criteriaare OK, and I think it's OK that vocab extensions are considered implementations, bur it's not clear how *extensions* serve to demonstrate the primary goal of demonstrated interop of documented features of PROV. But as I said, I think that will resolve itself when we look at the details of implementations

<zednik> Luc is breaking up over the audio

<Curt> Luc: there was a suggestion we should consider adding an example of bundles to FAQ

<Zakim> GK, you wanted to say the last example I saw was relating to the *previous* positon on namespace prefixes

<Curt> GK: the last example dealt with previous situations without nested identifiers

<Curt> pgroth: we clarified the way it worked, he wanted examples of using prefixes properly and how not to use them

<Curt> pgroth: need an action to add examples to FAQ

<Curt> Luc: I will do it (the example from his message?)

<Luc> http://lists.w3.org/Archives/Public/public-prov-comments/2012Oct/0019.html

<GK> ^^ not "without nested identifiers" but "without nested prefixes" - it's important to distinguish between these.

<Curt> Luc: the example he gave is valid, we need to explain why it is valid and add an example that is invalid

<pgroth> action Luc - add example of document/bundle to faq explaining validity

<trackbot> Created ACTION-122 - - add example of document/bundle to faq explaining validity [on Luc Moreau - due 2012-11-16].

<Curt> jcheney: agreed -- that's what we need to do.

<Zakim> GK, you wanted to ask about id used for bundle and entity within bundle

<Curt> Luc: we can close the issue now, with the coming action

<Curt> GK: you can have a bundle with an identifier, and use the identifier inside the bundle, to give provenance of the bundle itself. Is that ok?

<Curt> pgroth: that's a separate issue

<Curt> pgroth: that wouldn't change the spec

<Curt> GK: I thought you (paul) thought that would be invalid

<Curt> pgroth: delay considering that until later

<Curt> Luc: that is perfectly valid, and has an example in the DM

<Curt> zednik: the FAQ could attempt to address that

<Curt> pgroth: issue-569

<Curt> Luc: pending review waiting for james' response, came back to simon yesterday. he is happy with the suggestion, can close now

<Curt> pgroth: issue-475, mention

<Curt> pgroth: let's consider that (mention) at the end of this session so we can discuss it

<Curt> pgroth: editor review DM a final time for cleanliness/etc.

<Curt> Luc: how should we acknowledge reviewers?

<Curt> ivan: they will get listed as well as listing the working group

<Curt> ivan: put the same list of reviewers in each document

<pgroth> ACTION: Luc editor check [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action01]

<trackbot> Created ACTION-123 - Editor check [on Luc Moreau - due 2012-11-16].

<Curt> ivan: everything that needs to be changed has been changed?

<Curt> Luc: yes, except for final review, it is ready to go

<smiles> I'm trying to call in to the W3C bridge with code 7768 as said on the Wiki, but get "This pass code is not valid". Is there another code for today?

<GK> @smiles - that often happens to me … but usually works if I try again (i.e. re-enter the passcode).

<pgroth> we have now addressed all open issues (except mention) for prov-dm

prov-n

<Luc> http://www.w3.org/2011/prov/track/products/11

<smiles> * yes, I've tried a few times. not sure what the problem is, but will keep trying!

<Curt> Luc: last week, we agreed we would change scoping of prefixes, haven't received any feedback

<Curt> Luc: would be nice to have a few more examples

<pgroth> ACTION: Luc prov-n editor check [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action02]

<trackbot> Created ACTION-124 - Prov-n editor check [on Luc Moreau - due 2012-11-16].

<pgroth> smiles are you on

<pgroth> ?

<smiles> yes, but the sound keeps cutting in and out

<Curt> Luc: there is a typo in the current text

<Curt> Luc: all documents cross-reference each other, which URL should we use

<Curt> ivan: the dated URL

<Curt> ivan: it is a real pain, but they must always reference by the dated URI

<Curt> ivan: a global search/replace should take care of it.

<pgroth> ACTION: tlebo, jcheney, luc - check to see that all references refer to the dated documents (after a publication date is given) [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action03]

<trackbot> Sorry, couldn't find tlebo,. You can review and register nicknames at <http://www.w3.org/2011/prov/track/users>.

<Curt> Luc: we can't refer to those until we get the publication date

<Curt> Luc: is there a way to define the reference prefix up front and reuse it?

<Curt> ivan: (redacted)

<pgroth> w?

<pgroth> note: we are happy with prov-n

PROV-O

<pgroth> https://www.w3.org/2011/prov/track/products/3

<tlebo> http://www.w3.org/2011/prov/track/issues/552

<Curt> tlebo: issue-552, subclass, we did what they recommended

<Curt> tlebo: haven't heard back

<Curt> tlebo: we asked for a response on tuesday

<Curt> ivan: ok to close, we did what they suggested

<pgroth> ACTION: tlebo to add email link to the response page [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action04]

<trackbot> Created ACTION-125 - Add email link to the response page [on Timothy Lebo - due 2012-11-16].

<tlebo> http://www.w3.org/2011/prov/track/issues/592

<Curt> tlebo: he says terms are confusing, but his concern isn't clear

<Curt> tlebo: he expressed a concern, tim suggested an alternative approach, he hasn't responded to that

<Curt> laurent: wasInfluencedBy and wasInformedBy can get confused, there may be a better way to describe/depict their relationship

<Curt> tlebo: in the HTML is isn't as obvious which is the superproperty?

<Curt> laurent: yes, is isn't totally obvious in the HTML description of the ontology

<Zakim> tlebo, you wanted to ask hwo to make it "more top level" - it is already a superproperoty

<Curt> Luc: we changed the superclass description in the DM since Ralph reviewed, it might be more clear now

<Curt> Luc: Could revise the HTML description to clarify further

<Curt> jcheney: agreed, it says what we want it to say, but we might want to make it clear right up front which is the superproperty for querying and that you ought to use the more specific terms if possible

<Curt> ivan: might want to add the clarifying diagram

<Curt> pgroth: the document is already large, we are talking about ways to better guide how people should use the standard, but not affecting the standard itself

<Curt> pgroth: that sort of material, patterns, etc. should be in the FAQ

<Curt> ivan: we need to make sure those clarifications aren't lost, maybe include in the primer? where would people want to find that sort of material

<Curt> pgroth: I'm happy to have that added to the primer

<Curt> pgroth: that type of material -- I haven't seen that specific image or writeup

<Zakim> GK, you wanted to say - adding to primer means its fixed onpublication

<Curt> Luc: tlebo should forward Laurent's material to the list to consider for adding to the primer

<Curt> GK: the primer is fixed on publication, maybe link it to somewhere more dynamic

<Curt> pgroth: I like the FAQ for this type of stuff

<Curt> ivan: For usage patterns, I agree with GK, they will change/evolve, but the diagram from Laurent is more fixed

<Curt> GK: agreed, the diagram is different

<pgroth> ACTION: tlebo add a comment to use more specific things through document [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action05]

<trackbot> Created ACTION-126 - Add a comment to use more specific things through document [on Timothy Lebo - due 2012-11-16].

<Curt> tlebo: reassigned issue 592 to the primer

<pgroth> https://www.w3.org/2011/prov/track/products/10

<tlebo> http://www.w3.org/2011/prov/track/issues/461

<Curt> GK: difficult to follow cross-references when the document is printed

<Curt> Luc: in the DM, we numbered everything and refer by number instead of just the static link

<Curt> Luc: it was difficult to put in all those

<Curt> tlebo: :-)

<Curt> ivan: now is the time to make those sorts of changes

<Curt> tlebo: to address that, we would have a number for everything, and a table with all the numbers to index the terms

<Curt> tlebo: it may be difficult to do all that and not break anything

<Curt> tlebo: it is a purely editorial issue

<Curt> tlebo: if we can get through CR without that, then address it prior to next phase

<Curt> GK: this may be just too much work to implement

<pgroth> accepted: ISSUE-461 is editorial, the group agrees that this is ok to go ahead with CR and may look to address in the period of PR

<tlebo> http://www.w3.org/2011/prov/track/issues/593

<Curt> tlebo: need to change TTL example to exercise hadActivity

<Curt> tlebo: examples are considered editorial?

<Curt> ivan: yes, it is, but can it be done for CR?

<pgroth> ACTION: tlebo to add hadActivity example to prov-o [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action06]

<trackbot> Created ACTION-127 - Add hadActivity example to prov-o [on Timothy Lebo - due 2012-11-16].

<tlebo> http://www.w3.org/2011/prov/track/issues/479

<Curt> pgroth: we removed all TrIG?

<Curt> tlebo: there are a few remaining for 'mention'

<Curt> tlebo: reduced amount of TriG, and cited/described use of TriG

<Curt> ivan: clarify that all examples are informative

<Curt> ivan: must add that to the document

<Curt> ivan: then you can use TriG in examples and note that

<Curt> ivan: there may be a document from RDF group about TriG, and we could reference that later as an editorial change

<Curt> ivan: TriG reference must be informative, not normative

<Curt> ivan: it can reference it as a work in progress

<GK> http://www.w3.org/TR/2012/WD-prov-o-20120724/#Bundle -

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

<TomDN> http://aquarius.tw.rpi.edu/prov-wg/prov-o#Bundle

<Curt> pgroth: closing the issue, tim will clarify that examples are informative

<pgroth> ACTION: tlebo to add a statement on informative and normative in prov-o [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action07]

<trackbot> Created ACTION-128 - Add a statement on informative and normative in prov-o [on Timothy Lebo - due 2012-11-16].

<tlebo> http://www.w3.org/2011/prov/track/issues/566

<Curt> tlebo: fully addressed, waiting for daniel to respond

<Curt> tlebo: closing issue-566

<ivan> issue-491?

<trackbot> ISSUE-491 -- [external] feedback on prov:agent explanation. -- pending review

<trackbot> http://www.w3.org/2011/prov/track/issues/491

<Curt> tlebo: made some changes, Patrice likes it even less

<Curt> tlebo: doesn't like colloquial use of some terms and phrases

<Curt> tlebo: wants things expressed in logic terms

<Curt> pgroth: his phrasing would rewrite the document in a rule based form

<lebot> hello?

<Curt> ivan (and others): the proposed language is very convoluted for people to read, we shouldn't do it

sorry about the noise

<Curt> ivan: some of the wording could be better

<Zakim> GK, you wanted to suggest s/used by/used with/

<Curt> GK: change "used by" to "used with"

<Curt> ivan: yes, that may be a simple way to address some concerns

<Curt> pgroth: are these in many places?

<Curt> tlebo: I removed some of the objectionable language

<Curt> ivan: why was he even more upset?

<Curt> tlebo: we were reusing prov:AgentInfluence, but we change our usage of that, with a better definition

<Curt> tlebo: we've addressed some of the expressed concerns

<Curt> tlebo: I think we've addressed it all

<Curt> pgroth: we don't want to use the proposed phrasing, I think this has been adequately addressed

<Curt> tlebo: closing issue 491

<Curt> https://www.w3.org/2011/prov/track/actions/116

<Curt> Luc: Tim will address action 116 post-CR release, determine if it is doable

<Curt> pgroth: Tim will do an editor check of PROV-O

<pgroth> ACTION: tlebo editor check prov-o [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action08]

<trackbot> Created ACTION-129 - Editor check prov-o [on Timothy Lebo - due 2012-11-16].

<pgroth> very happy with prov-o

http://www.w3.org/2011/prov/track/products/12

PROV-Constraints

<Curt> pgroth: All issues have been addressed, sent back to reviewer

<Curt> jcheney: he has had a week to consider our responses

<Curt> ivan: were any of the resolutions controversial?

<Curt> jcheney: there were a few common themes, some were simply typo/rewording

<pgroth> close ISSUE-587

<trackbot> ISSUE-587 Concerns about analogies to RDF blank nodes/semantics closed

<Curt> group: (we like tracker!)

<pgroth> close ISSUE-586

<trackbot> ISSUE-586 The description of 'toplevel bundle' as 'set of statements not appearing in a named bundle' is unclear closed

<pgroth> close ISSUE-582

<trackbot> ISSUE-582 'of their respective documents.' should be '... of their respective instances.' closed

<Curt> jcheney: some of the suggestions might be more appropriately addressed in the semantics document

<Curt> jcheney: they didn't fit the nature of the the constraints goals

<Curt> ivan: maybe we didn't clarify the goals of the document?

<Curt> jcheney: I tried to elaborate purpose of document, that somewhat addresses that concern

<Curt> pgroth: current description of constraints document

<TomDN> I think this sentence addresses a lot of his concerns as well: "Further discussion of the semantics of PROV statements, which justifies the definitions, inferences and constraints, and relates the procedural specification approach taken here to a declarative specification, can be found in the formal semantics [PROV-SEM]. "

<Curt> Luc: message in document is fairly clear what we intend for the document

<Curt> ivan: that description sounds ok, need to be clear that this is a precise way to check validity of PROV

<Curt> ivan: Antoine may be looking for semantics -- that isn't the goal of this document

<Curt> jcheney: that is how I have addressed the issues

<Curt> pgroth: add 1 sentence to description on constraints document -- this defines a precise way to validate provenance

<pgroth> This document defines how to precisely validate provenance documents.

<Curt> jcheney: will add that sentence

<Curt> pgroth: I read all the issue responses and thought they were good -- so did luc

<Curt> jcheney: issue-585, described why things are worded the way they are

<pgroth> close ISSUE-585

<trackbot> ISSUE-585 Suggestion to avoid discussing how to 'apply' constraints; clarify what it means to 'satisfy' constraints closed

<Curt> issue 576, the term merging was replaced with unification that is more accurate

<pgroth> close ISSUE-584

<trackbot> ISSUE-584 The nonstandard/procedurally defined 'merging' operation on terms closed

<Curt> ^576^584

<Curt> jcheney: issue 583, rewrote wording of equivalent instances

<pgroth> close ISSUE-583

<trackbot> ISSUE-583 Questions concerning what it means for applications to treat equivalent instances 'in the same way', particularly in bundles. closed

<Curt> jcheney: issue 581 wording around normalization/equivalence

<Curt> GK: equivalence is really observed behavior -- given the same situation, you should get the same provenance

<Curt> jcheney: I'll reword some of this and circulate for comment

<pgroth> ACTION: jcheney to add a bit of text around equivalence and remove normative SHOULD [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action09]

<trackbot> Created ACTION-130 - Add a bit of text around equivalence and remove normative SHOULD [on James Cheney - due 2012-11-16].

<GK> ^^ Not "equivalence", but "treat in tghe same way" is what is observed/able behavious.

<Curt> issue 581, we agree we are not specifying the algorithm, will clarify,

<pgroth> close ISSUE-581

<trackbot> ISSUE-581 Suggestion to avoid wording that 'almost requires' using normalization to implement constraints closed

<Curt> jcheney: issue 580, definitions for expanding compact language not needed; response -- yes, we do need to define how those things work

<pgroth> close ISSUE-580

<trackbot> ISSUE-580 Suggestion to drop definitions in section 4.1 since they are not needed if the semantics is defined more abstractly closed

<TomDN> issue-578?

<trackbot> ISSUE-578 -- Use of "equivalent" incompatible with common uses of the term in logic/mathematics -- pending review

<trackbot> http://www.w3.org/2011/prov/track/issues/578

<Curt> jcheney: issue 578, we defined equivalence only on valid documents, not arbitrary documents

<Curt> jcheney: we need to consider equivalence for other scenarios beyond validity

<pgroth> close ISSUE-578

<trackbot> ISSUE-578 Use of "equivalent" incompatible with common uses of the term in logic/mathematics closed

<Curt> ivan: for the purpose of this document, our description is sufficient

<Curt> jcheney: yes, once we clarify the purpose of our document, the concern becomes somewhat moot

<TomDN> issue-577?

<trackbot> ISSUE-577 -- Terminology: valid vs. consistent -- pending review

<trackbot> http://www.w3.org/2011/prov/track/issues/577

<Curt> issue 577, we use the word "valid" where logic uses "consistent",

<Curt> ivan: this document isn't meant for logicians

<pgroth> close ISSUE-577

<trackbot> ISSUE-577 Terminology: valid vs. consistent closed

<Curt> jcheney: we are using the words appropriate for our purpose

<pgroth> close ISSUE-576

<trackbot> ISSUE-576 logical definition and comments on prov-constratins closed

<Curt> issue 556, translating constraints to prov-o out of scope

<Curt> pgroth: that is a concern of implementers

<pgroth> close ISSUE-556

<trackbot> ISSUE-556 public comment: should qualfied and unqualified versions the same closed

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

<pgroth> ACTION: jcheney editorial check on prov-constraints [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action10]

<trackbot> Created ACTION-131 - Editorial check on prov-constraints [on James Cheney - due 2012-11-16].

<pgroth> ACTION: jcheney add response email to responses to public comments page [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action11]

<trackbot> Created ACTION-132 - Add response email to responses to public comments page [on James Cheney - due 2012-11-16].

<pgroth> we are happy with constraints

<pgroth> 15 minute break

<pgroth> start at 11

<lebot> i added a comment to https://www.w3.org/2011/prov/track/actions/125 can I close it?

Mention

<pgroth> Scribe: Tim Lebo

<pgroth> starting again

<lebot> paul: we came to a simple definition of mention, from many before it.

<lebot> … connects Entity in one bundle to an Entity in another bundle. It's a kind of specialization

<lebot> … Luc's response to Graham's public comment

<lebot> … "at risk" is not appropriate for mention.

<lebot> … having "at risk" in CR - does not look good.

<lebot> … need to settle it now. Make it lean.

<lebot> ivan: at CR, "at risk" is one that the WG thinks it has an issue implementing. But mention is not an implementation issue, it's a design issue.

<lebot> … if design, then it is an abuse of "at risk"

<lebot> pau: the chairs do not want to abuse "at risk".

<lebot> … thus, include or exclude now.

<Luc> @lebot: can you use pgroth as handle?

<lebot> … we've spent a LOT of time on mention. we need to go from that work.

<lebot> pgroth: lets hear case against as it stands.

<lebot> … does anybody want it in?

<lebot> … who wants it out?

<lebot> … we'll decide in or now today.

<lebot> GK: debate has been going on for long time.

<lebot> … we can't conflate previous things with what it is now.

<lebot> … feel there is an attempt to introduce something which cannot be specified in RDF.

<lebot> … BUT the public objection is NOT ^^^

<lebot> … basically, I don't know what it is trying to say.

<lebot> … what does it mean?

<lebot> … what is new beyond what we already have?

<TomDN> (original email: http://lists.w3.org/Archives/Public/public-prov-comments/2012Aug/0001.html )

<lebot> … my claim is that it does not add anything.

<lebot> pgroth: who wants, will use mention?

<lebot> jcheney: at last F2F we discussed this.

<lebot> … strong motivation in ontology to relate MentionOf relation two two entities.

<lebot> (asInBundle)

<lebot> … the idea is to translate mention of DM into two triples in RDF.

<lebot> … how to convert when round tripping DM PROVO DM?

<lebot> … what if two mention triples?

<lebot> … you'll get confusion when coming back to DM.

<lebot> (The "limitation" is that you an only be asInBundle to one bundle)

<lebot> … seems like a misalignment in the serializations.

<Luc> http://www.w3.org/TR/prov-constraints/#unique-mention

<lebot> … could be viewed as doing different things in PROVO and DM.

<lebot> luc: we introduced the constraint the mention must be unique - so you can't have the confusion that jcheney suggests.

<lebot> lebot: I'm happy with it.

<Luc> specialization is not reflexive, so they must be different URIs

<lebot> lebot: when we're trying to interconnect descriptions of entities in others' bundles, it's a natural thing to do.

<lebot> ivan: do you use the same URI?

<lebot> lebot: you can do either, depending on what you want to do.

<lebot> Curt: mention is the only capability to reference into the bundle. You'll run into problems if you don't have it.

<lebot> TomDN: i support using mention of.

<lebot> … a lab with multiple documents and multiple people. You just want to mention it, not repeat the provenance.

<lebot> … it's interesting to provide your own view on the entity that you're using.

<lebot> pgroth: we have specialization and alternate of.

<Luc> In view of implementation phase, can we see who will make use of the mention construct in their implementation?

<lebot> … the key aspect of mention of is that you name the entity and the bundle in which the entity is described. The Bundle IS the specialization.

<lebot> … without mention, you can still link the entities, but you lose the ability to mention the bundle.

<lebot> +1

<lebot> Luc: who will implement it?

<lebot> TomDN: we will.

<lebot> hook: mentionOf, but used unique identifiers to link across. didn't use mentionof

<lebot> … trying to link bundles. it was easier to not use mentionOf.

<lebot> hook: KISS philosophy.

<Zakim> lebot, you wanted to state that the system hook is using is one system, not multple

<lebot> lebot: mentionOf's power comes in when you don't have control over the entire system.

<Curt> +1 lebot

<lebot> hook: we should force people to use mentionOf to increase interoperability.

<lebot> pgroth: we can't force people to use it (and shouldn't)

<lebot> … we should offer it for people to use.

<lebot> hook: sounds like it doesn't hurt to leave it in, helps to connect.

<lebot> +1 hook

<lebot> Luc: [not?] concerned with comments that Graham raises.

<lebot> … but the doubt is if it is really useful or not.

<lebot> … believe in stitching histories.

<lebot> … we need a construct for it.

<lebot> … BUT concerned if it is a subtype of specialization.

<lebot> … working to develop the use cases.

<lebot> … as a sub property of specialization, the lifetimes are maintained.

<lebot> … in the use case, the timeline constraint may not apply.

<TomDN> +q

<lebot> GK: not sure if it breaks specilization

<lebot> (+1 to GK)

<lebot> luc: unsure about making it a type of specialization.

<lebot> … we're stuck with keeping mentionOf as specialization (and not alternate)

<lebot> … if it's specialization, does it break?

<lebot> (-1 that it's broken as specialization. It's inherently specialization)

<lebot> TomDN: how does it break as specialization?

<lebot> … did we want the validity over different bundles?

<lebot> … at what point do we make a new entity?

<lebot> (+1 Tom)

<Paolo> I missed all of Tom's comment -- low voice

<Paolo> ok thanks

<lebot> pgroth: the question: do we have validity over different bundles

<lebot> TomDN: luc's problem goes away once the entity is in a different instance.

<lebot> … entity in a different instance, valid, same instance different bundle = invalid

<lebot> … generation and invaliation of both entities, specialization applies and must have a lifetime.

<lebot> TomDN: impossible to make valid if repeating the mention?

<lebot> pgroth: it done'st make it invalid, but …. (?)

<lebot> jcheney: inférences on uniqueness are flagged as at risk.

<lebot> … if somebody is at risk, we can decided to remove it w/o going to LC

<lebot> @luc, you're abusing mention of for the wrong use cases. (it appears)

<lebot> jcheney: is it possible to take out parts of the at risk?

<lebot> ivan: mention is a design feature, defined [as specialization]. it is a design element.

<lebot> … it is all or nothing.

<lebot> jcheney: we can remove it all. If we change it, then it's a design change.

<lebot> GK: can't you drop parts of the definition and not others, providing that the others are not changed?

<lebot> ivan: feature at risk, feature defined. Remove or keep it.

<lebot> … splitting hairs is sticky.

<lebot> Curt: I don't follow the issue. It DOES fit into specialization.

<lebot> … as a primary producer, I wont' use mention of, but for anyone that wants to augment my Entiteis, they need mentionOf to do it.

<Luc> @tlebo, can you clarify why i am abusing it?

<lebot> … the third party needs it.

<lebot> @luc, I'm not clear on what you're trying to do, but it doesn't sound like mentionOf

<lebot> Curt: when yoiu do your own provenance, you ond't need it, but metnionOf lets you "reach into" someone else's bundle.

<lebot> jcheney: second order provenance and linking.

<lebot> … but it's also true for other things.

<lebot> … are we solving a specific problem and not the more general?

<lebot> … it's clear that there is a need, but is it justified?

<Curt> entity is pretty much our most general thing to refer to

<lebot> … I am still uncomfortable with mentionOf

<lebot> … if it was lightweight with no inferences, then fine. But we might get into trouble later.

<lebot> … as things are, it doesn't seem like we should kill it, but people might trip over it later.

<lebot> hook: the linking of bundles should be in the model, we should not rely on a serialization

<lebot> @hook how are they different?

<TomDN> +q

<lebot> pgroth: there are existing ways to annotate. Refer to things an annotate them.

<lebot> … open annotation

<lebot> … some let you point to named graphs.

<lebot> …. well out side of our scope.

<lebot> … but those things are not for provenance.

<smiles> So mentionOf is just a way to reference a part of a document without reference to the serialisation format? Is mentionOf really to do with provenance apart from being arbitrarily restricted to PROV?

<lebot> … open annotation is not a standard, but is in w3c

<lebot> hook: having it formally in DM would uniformly manifest implementations in different encodigns. we're not relying on serializations to do the linking.

<lebot> pgroth: right now, you can use RDF linking.

<lebot> TomDN: should we drop it and put it into a note?

<lebot> … here is how to link" in FAQ...

<lebot> … we can change as we see fit.

<lebot> GK: in IETF, "experimental track", mention of is in this.

<lebot> … best we can do is to put FAQ

<lebot> ivan: it is a nice idea.

<lebot> … we have notes, we'd just be adding one more.

<lebot> pgroth: if that's what we want to do, it'd go AQ

<lebot> … we can't start a new note

<lebot> ivan: agree with graham that AQ is to locate provenance of a given resoruce.

<lebot> … that's different than mentionOf

<lebot> … it doesn't fit

<lebot> hook: how many use cases involve mentionOf?

<lebot> … for what we do, it would be useful.

<lebot> Curt: the key is not provenance expression/represtionation, ti's for analysis.

<lebot> GK: how important is interoperability at the analysis/

<lebot> ?

<lebot> hook: it is very important.

<lebot> … each bundle is handled by different institutions, gov entities.

<lebot> … interop is key here.

<pgroth> who just joined/

<pgroth> ?

<lebot> Curt: we have a lot of cases where data is processed, then next org processes. each uses their own bundles.

<lebot> … each needs a way to reference across those bundles.

<lebot> … seems that mentionOf provides a capability that will be needed at some point.

<lebot> Luc: jcheney, you'd be more comfortable to get rid of the inference?

<lebot> jcheney: uniqueness constraint makes to align with provo round tripping.

<lebot> … it's not clear that it buys you much.

<lebot> … you could just state the specialization.

<lebot> (I think the 'you don't get anything" assumes that you "have it all" and does not consider the practicality of the problem)

<lebot> jcheney: not hearing strong objections, but nobody is giving specific uses for it (?)

<Paolo> +1 for unlinking MentionOf from Specialization (if I understand James correcty)

<lebot> jcheney: not worth rolling all of it back

<lebot> Luc: we didn't want to make it a top-level, that's where we started.

<lebot> jcheney: not worth blowing the whole thing up over.

<Luc> is there opposition to remove it?

<lebot> pgroth: straw poll on mentionOf

<lebot> (this will decide who I sit with at lunch, btw)

<lebot> SamCoppens: selective removal okay?

<lebot> pgroth: no, since it changes the spec too much.

<pgroth> straw poll: who objects to keeping mentionOf?

<GK> +1

<smiles> +1

<Paolo> 0

<ivan> 0

<khalidBelhajjame> 0

<pgroth> straw poll: who objects to removing mentionOf?

<GK> 0

<khalidBelhajjame> 0

<lebot> :-(

<TomDN> 0

<zednik> 0

<SamCoppens> 0

<hook> +1

<Paolo> 0

<lebot> GK: I would formally object in its current form.

<smiles> I would not formally object. I was indicating that I think it is better not to be in the spec in the straw poll.

<lebot> Curt: I think it's valuable, but I won't formally object.

<GK> Longer response, in IRC for lack of time:

<GK> - yes, there are valid use cases, strong motivation

<GK> - I don't recognise them in the mentionOf as described (my complaint) in a way that can't be done without mentionOf

<GK> - some of those use-cases don't map to present-day RDF semantics - I worry about this, as we'd end up building on sand if we try to impose these semantics

<GK> - not defining it now doesn't mean it can't be defined later

<Paolo> may be back later

<khalidBelhajjame> I may be back later

<lebot> tlebo: If GK's formal objection is the thing to scare away this construct, then I'd be willing to bring RPI's formal objection to dropping it.

<lebot> … but this is weighted by the fact that I'm exhausted with supporting this construct.

<lebot> ivan: formal objection is a HUGE thing.

<pgroth> start again in one hour

<smiles> OK thanks

mention of and CR

<smiles> My objecytion was not formal

<pgroth> scribe: James Cheney

pgroth: 30 minutes on mention
... have formal objections changed?

<ivan> scribenick: jcheney

GK: after lunch discusion with tlebo
... thinks problem may be fixable with changes to descriptive text, but not sure yet

ivan: can we do it now?

gk: maybe not enough time
... can we proceed on assumption it will be fine?

luc: wants certainty
... can we take an hour and do it now?

GK: will look at it offline now.

PROV-XML

pgroth: Graham will look at document for ~1hr, we move on to prov-xml, goal is to come back to CR vote today

[luc is chair]

Luc: prov-xml was reviewed over past week (James, Paul, Luc)

would like to decide on release as fpwd

scribe: would like to decide on release as fpwd

zednik: document mostly content complete, adding bundles today
... should be finished in ~5min
... reviews identified typos & rephrasing, had some questions about design/descriptions
... discussion topic list to respond & discuss feedback
... most feedback has been incorporated
... all 3 said it was ok to proced to fpwd
... currently addressing more complex identifier issues

curt: also thinks things are OK

smiles: wanted to point out comment that might have been missed
... delegation element in prov-xml: schema description is different from actual schema
... but also agree document is ready for release

<ivan>

zednik: will double check

pgroth: do we vote next or have content discussion?

Luc: discuss reviews and any tecnical issues first, then vote

<smiles> @zednik: the issue was that the activity was an option of actedOnBehalfOf in the schema, compulsory in the schema fragment in the HTML

pgroth: thinks its OK for FPWD, would like to discuss technical issues

curt: would like to discuss 572

<TomDN> issue-572?

<trackbot> ISSUE-572 -- What constraints should we have on ordering of elements within the main complexTypes? -- raised

<trackbot> http://www.w3.org/2011/prov/track/issues/572

jcheney: mostly happy, can discuss offline

pgroth: also wanted to suggest signposting/context, is this intended before fpwd?
... meaning expanation of the style of schema being used (salami slice pattern, etc)

ivan: sounds good, helpful to reader

zednik: prov-xml group is discussing adding a design section, explain salami slice pattern, not sure if it will go in before fpwd

<pgroth> i wouldn't want it to delay fpwd

Luc: conform happy with document release, flagged some technical issues
... need to catch up on mailing list traffic, but OK with flagging as outstanding issues in text as notes
... to avoid giving impression that it is a final design
... design section sounds useful
... timetable to release: need not be ASAP, but would be good to sync with CR
... to give time to write section

<TomDN> +1 for synchronous release

pgroth: would like it to be released synchronously with CR/primer, etc.
... have gotten burned before by piecemeal release
... prov is the family, would like releasing as such
... no rush to get xml out, but there are minor things we can do to improve accessibility

ivan: we clearly don't have enough documents to publish, so let's add one
... owl WG had relatively short overview document published with rest
... otherwise family of documents becomes messy

Luc: not committed to it in charter extension, avoid overcommitment

ivan: together with CR release?

Luc: not enough time

<Curt> copy the intro from the DM

<pgroth> ACTION: pgroth to draft a first one page overview [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action12]

<trackbot> Created ACTION-133 - Draft a first one page overview [on Paul Groth - due 2012-11-16].

pgroth: will try to draft 1 page, group will look at it. as curt says, this is already done in most documents

luc: can reuse presentation tutorial materials.
... informal poll to gauge positions on fpwd
... is ther opposition to prov-xml fpwd release?

<Paolo> no objection

[crickets chirping]

<smiles> no objection

sorry

Luc: what do we want to finalize before fpwd?

pgroth: want 1 para about design + "warning, this is a fpwd, subject to change"

Luc: any other input?
... can we confirm prov-xml as short name?

<Luc> proposed: To release prov-xml as a first public working draft, after adding design overview and sign-posting issues under consideration, with prov-xml as short-name

<TomDN> +1

<ivan> +1

<pgroth> +1

<smiles> +1

<Curt> +1

<SamCoppens> +1

+1 UoE

<lebot> +1

<Luc> accepted: To release prov-xml as a first public working draft, after adding design overview and sign-posting issues under consideration, with prov-xml as short-name

Luc: now have time to discuss technical issues

<TomDN> issue-572?

<trackbot> ISSUE-572 -- What constraints should we have on ordering of elements within the main complexTypes? -- raised

<trackbot> http://www.w3.org/2011/prov/track/issues/572

<ivan> issue-572?

<trackbot> ISSUE-572 -- What constraints should we have on ordering of elements within the main complexTypes? -- raised

<trackbot> http://www.w3.org/2011/prov/track/issues/572

Curt: Mapping from PROV-N to PROV-DM into xml schema decided to keep same order of sub-elements as in prov-dm
... Current rationale: atributes are ids
... ordering of content is static matching prov-n
... except for optional attributes which are unordered

<pgroth> wonder why there's no issue about sub typing?

Curt: could relax ordering, or require ordering of attributes
... Concern that ordering makes it easier for processing, but harder for generation
... unlike prov-n

jcheney: happy with wat it is, decreases tax on everyone to normalize
... happy with way it is, decreases tax on everyone to normalize

luc: had idea to require prov attributes to appear first, then non-prov
... use xsd:any for all the rest
... should make it easier to convert between xml and other PL embeddings
... with xml, thinking about serializations but also queries
... does order have impact?

jcheney: probably XQuery with unordered xpath axes is enough, so order probably not a big issue for queries

pgroth: not sure of issue

luc: orm will want to be able to find prov:type
... so mapping will be challenging

jcheney: we don't need to solve this now necessarily

ivan: can ask for feedback

pgroth: automated generation tools are a use case, we should flag this for asking for feedback

<Zakim> pgroth, you wanted to say we should test

luc: issue remains open, but will be signposted

pgroth: wants to discuss subtyping

<Zakim> pgroth, you wanted to ask for about sub typing

pgroth: if you look at prov-xml, many subtypes are defined through use of prov:type
... in prov-o, a revision has a corresponding relation
... why can't xml / xsd do something similar

curt: also would like to do this

zednik: followed prov-n initially, but can explore and add in after fpwd. note in each section to explain this

pgroth: raise issue?

zednik: did look at subtyping early, but mainly entity and agent and it didn't seem to gain a lot since these subtypes don't have additional elements/attributes
... but relations may have a benefit

pgroth: in xml, you see agent but not person etc.
... writing xpath query to ask for people is easier if the element name is prov:person

zednik: would have to specialize complex type and add new toplevel element referencing it
... this should work, but hasn't been tried yet. may work for entity and agent subtypes too.

Luc: will have to add subtype and new elements. don't we want to allow use of person, etc. wherever an agent is allowed?
... but then haven't you fixed all the subclasses of entity/agent, forbidding extensions?

zednik: not familiar with extended types in xml, but should allow specialization / subtypes without using substitution groups

Luc: something to keep in mind when looking at revised design.

<pgroth> did someone raise the issue?

zednik: suggest we mark the terms that use prov:type for subtyping as something that might change

<pgroth> issue: prov-xml subtyping needs to be marked in the document

<trackbot> Created ISSUE-595 - Prov-xml subtyping needs to be marked in the document ; please complete additional details at http://www.w3.org/2011/prov/track/issues/595/edit .

pgroth: whoa!

luc: next issue, identifiers/qnames

<Luc> entity(ex:0001)

luc: can write entities like this

ivan: this is why rdfa does not use qnames

luc: grammar accepts qualified names but xml schema requires qnames

ivan: [shrug] life sucks

luc: can define new type of strings that match this
... in prov toolbox, using in non validating mode so these recognize as qualified names but painful

zednik: should try to determine what is best for xml to use as identifier
... identifying scheme for prov-n makes sense in rdf, may not make sense in xml
... defining our own string subtype may not be best either

pgroth: agrees with stefan's approach. made prov-n open-ended for human consumtion
... with xml, need to be more restrictive to remain compatble for tools, even if it constraints what you can use as ids

tlebo: rdf/xml has same problem,

<Luc> my concern is that people will generate xml that does not validate

pgroth: design for tooling

ivan: it is a choice to allow more liberal strings, but will not work well with tools

<lebot> +1 pgroth and zednik on letting prov-xml constrain, c.f. prov-o's "type" must be a Resource and not Literals, as prov-n permits.

pgroth: does qname resolve to uri? main serializations will be xml, rdf/turtle
... we don't have to define in documents, but should say somewhere what subset of ids are interoperable across main formats.
... "don't do this"

luc: concerned people will generate xml serializations that don't validate because of ids
... qnames are very restrictvie

curt: seems ok to say "if you want to interoperate, do this"

hook: no xlinking

pgroth: shouldn't define our own ids. do people use something other than qnames?

laurent: people used to use urn, now uri/url

<zednik> +1 pgroth for determining what is best for ids from xml community, and use that

ivan: there are organizations whose internal identification of items is similar, rdfa discussion began because news organization wanted to use similar names
... rdfa avoided use of qnames

pgroth: also allowed in prov-o, prov-n

ivan: defining new id type worse because many xml tools assume id attribute is of a specific form (?)

Luc: we use prov;id, not toplevel id

ivan: some tools recognize/exploit atributes declared

<ivan> s/atributes/attributes/

jcheney: will ask ht

<Curt> This could be an explicit question for FPWD review too

luc: prov-dm uses qualified names as shortcut for uri
... can reconstruct full uri
... not done in xml by default

<Paolo> I will have to go soon -- are you planning to discuss prov-dictionary next?

luc: we need to state the convention

<pgroth> @paolo yes

luc: plan: flag issue, have james ask henry thompson

zednik: wanted to add that we could put forth question + possible direction such as xsd:anyURI

luc: may lose some benefit of xml?
... congratulations to prov-xml team

prov-dictionary

luc: renamed collections to dictionaries, then decided to remove from dm leaving lean collections
... decided to create note for dictionaries, starting with all text from older verisons of prov-dm/prov-o
... but some work is needed. who will work on it?
... comments?

<TomDN> +q

<Paolo> +q

TomDN: what is timetable?

luc: to be detemined

TomDN: synchronous release?

Luc: no, later than cr release
... but before end of wg
... including time for iterations

Paolo: discussed earlier, and when we decided on note, ownership was assigned to stian with paolo agreeing to help
... but was involved in other documents so did not have time
... talked with stian and discussed timetable but this hasn't been realized
... plan to ask stian if interested, volunteer to help, otherwise try to pick up
... would still like to see it happen
... should be able to start spending time on it after holidays

Luc: can you really do it?
... in terms of bandwidth

Paolo: will have more in January, not before

<lebot> @Paolo , we all have more bandwidth later. Until we don't ;-)

Paolo: can make time for it
... don't think we're too far

<SamCoppens> Tom and I would volunteer to help with the note

<lebot> good point, it was carried to Last Call drafts :-/

Paolo: material in note is not starting from scratch

pgroth: timetable would like to see fpwd or new release on notes before holidays for all documents

<lebot> +1 to a FPWD for collections before xmas

pgroth: there on most things already (prov-aq, prov-dc)
... collections needs editorial work beyone existing content

Luc: at f2f3 took out of rec track document, no activity since then
... if someone volunteers to work on it before holidays, great, if not, we may not have time to finish it by march

tlebo: reinforcing paolo's comments: content is from pre-last call
... can support with prov-o parts

paolo: will struggle between now and end of year but can try to make time
... spike in teaching activity now

<pgroth> @paolo that's why we need something else

paolo: unlikely to find more than 1-2 days
... was assigned to stian, so begin by checking whether he still plans to do this

sam: tom and i will definitely help, could take lead if needed

<Paolo> excellent I would definitely help out

luc: sounds good!

pgroth: stian may be busy, so extra help would be good; stian is a core implementor in taverna, & working with open annotation
... implementations more important

<pgroth> ls

<lebot> +1 drink each to @SamCoppens and @TomDN this evening ;-)

luc: NB: christmas is only ~6 weeks away

(oops that was me) NB: christmas is only ~6 weeks away

<lebot> @Paolo can we wrap our arms around the raw materials?

<Luc> ACTION: SamCoppens to draft a timetable for prov-dictionary for the next teleconference [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action13]

<trackbot> Sorry, couldn't find SamCoppens. You can review and register nicknames at <http://www.w3.org/2011/prov/track/users>.

@tlebo http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dictionary.html

<Paolo> cool, have to go bye everyone

luc: completed prov-xml, prov-dictionary
... allocate 30-minutes to prov-sem?

<Luc> scribe: TomDN

prov-sem

<jcheney> http://www.w3.org/2011/prov/wiki/FormalSemanticsLC

<pgroth> ACTION: TomDN draft a timetable for prov-dictionary for the next teleconference [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action14]

<trackbot> Created ACTION-134 - draft a timetable for prov-dictionary for the next teleconference [on Tom De Nies - due 2012-11-16].

jcheney: Update on PROV-SEM.

<lebot> @SamCoppens http://dvcs.w3.org/hg/prov/file/tip/ontology/prov-dictionary.owl

jcheney: Most of what's here: http://www.w3.org/2011/prov/wiki/FormalSemanticsLC is aligned with the LC docs
... With the CONSTRAINTS, we only enable people to track the constraints. But with SEM we could formalize all that more cleanly and acceptable for logics people
... It's kinda hard to write that stuff down in HTML, instead of in for example LaTeX
... There's an old Latex->HTML tool, but it's not conforming to the recent standards
... If I could write it as Latex, producing this note would be easier

Luc: What's your sense of timetable?
... And are there people who could help you?

jcheney: Help would be good.
... Now is a good time for me to do it.
... But time that I wanted to spend on this has gone to the constraints.
... I could definitely use people that can do the math markup

pgroth: Go ahead an focus on the content, and we can see if we can find people to make it look nice

ivan: it could be on the wiki after the WG closes. Then it has a URI and is read-only

jcheney: But is that OK for a formal Note?

ivan: no.

<zednik> http://gva.noekeon.org/blahtexml/

<lebot> @zednik what a great name for a tool.

Luc: Isn't there a tool at W3C to turn a wikipage into a note?

ivan: Sandro had some python tools, but I don't know whether that would work. You'd have to ask Sandro.

TomDN: I think the content is most important, to address comments about the semantics. let's focus on that first

jcheney: I think a lot of people thought it'd be nice to have this, so it's definitely worth doing. The feedback was useful, but not the main reason to produce the Note

pgroth: Conclusion: James keeps working on this in the way that's easiest for him, and then someone looks at the presentation stuff later.

Luc: Timetable?

jcheney: I need about a week (continuous) work on this.
... The week of the holidays seems reasonable for a first draft

<jcheney> http://www.w3.org/2011/prov/wiki/FormalSemanticsLC#Inferences

jcheney: If you look at the end of the document, you'll see that I've already converted most stuff into the subset of LateX that Wiki supports.

<pgroth> 30 prov logic parsers

<pgroth> all independent implementations

<pgroth> we are happy

everyone: We are all happy

pgroth: Is there anyone on the phone that has comments on anything on the agenda?

Mention

Luc: Graham has thought about Mention.

GK: I think I have an explanation of it that I'm OK with.
... I hope it aligns with what is meant in the document.

Luc: So you're not proposing a change of design, but a textual change?

GK: Yes, it's an explanatory change.

(Taking a break until 3:15 )

<smiles> OK, talk then

<laurent> @jcheney Instructions to export wiki pages to HTML used for SSN http://www.w3.org/2005/Incubator/ssn/wiki/Publishing_Incubator_Group_Documents#Export_into_HTML_from_MediaWiki

(and we're back ! )

pgroth: Graham to propose editorial changes

<GK> http://www.w3.org/2011/prov/wiki/User:Gklyne

(who just joined? )

GK: Please see the link for the text regarding my suggestions

<jcheney> can someone resend link

GK: This is based on the description of Mention in PROV-DM

<SamCoppens> http://www.w3.org/2011/prov/wiki/User:Gklyne

<jcheney> @Sam thanks!

GK: I think this way, the examples could be done without TRiG

<pgroth> ace Luc

Luc: In your first 2 sentences, you talk about the same entity using the same name, but with different descriptions in different bundles
... However, we can't have the same name
... Why do you have to indicate that they have the same name? Why not just the same entities with different descriptions?

GK: Valid point, I was just working from a specific use case
... It may not be necessary in the eventual descriptive text

Luc: are you introducing a new inference?

GK: I don't think so.

<lebot> I pause on "the descriptions may be based on observations of different specializations "

GK: To be clear, these are my thoughts on the matter, not something that should go directly into the description

<lebot> it seems to impose a specialization of an entity every time someone attempts to fix an aspect of the entity.

luc: Is "An application may have access to additional out of band information " there to explain the difference with /just/ a specialization?

GK: yes

<lebot> I pause on "about the specialization of e1 that is described in bundle b" since a specialization is not asserted - e1 is itself!

Luc: Example: ratings. If I rate something that lasts an hour fast, someone else might rate it differently

<lebot> I very much like "The mentionOf construct provides a way to introduce a new entity that is the basis for observations in a specified bundle"

GK: I think we're talking about the same thing

Luc: Do you want to add these inferences to the document?

GK: No, they are to help capture the essence of the text

Luc: Does this mean that you are now happy with mention? (If we do these edits in the text)

GK: I'd say yes, if my interpretation is what's meant in the document

Luc: Attempting to assess the changes to be made

<pgroth> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#term-mention

<lebot> I think the term "mentionedIn" is too broad from what we currently have: "asInBundle"

<jcheney> "introduce a new entity that is the basis" -> "relate an entity in the current instance to another one that is the basis..." ??

GK: I had trouble with "Some applications may want to interpret this entity e1 with respect to the descriptions found in the bundle b it occurs in."

Luc: Yes, it looks liek we actually mean "The description of the entity in bundle b"

s/liek/like

GK: also, "additional aspects"

tlebo: but "aspects" is central in the definitions of alternate and specialization

<lebot> Central to mention: "The primary author did not see fit to specialize, but the secondary consumer/author *does* see fit to specialize the entity".

Luc: We didn't want a formal definition of aspect

tlebo: the term "additional aspect" just refers to the specifying of the bundle. After that you can add whatever you want

GK: My problem is that it's focusing on the mention as the aspect

Luc: I want to know exactly which edits we want to make

GK: I was treating this as trying to capture the same information as in the document
... as a replacement

smiles: I personally find the original text clearer than Graham's
... To me, it doesn't seem to be about provenance, and not useful.
... I wouldn't formally object, but I wouldn't use it
... everything else in the document describes things in the past. But this doesn't.
... So it's not really provenance

scribe: I don't see a problem with it being at risk. But since we want it in or out now, I would vote for out

<Curt> It allows us to tie additional information to provenance information

TomDN: But alternate and specialization technically don't describe things in the past, so why block mention for that reason?

smiles: because they do describe "this thing was alternate of this thing" in the past

TomDN: I think mention does that as well, just with a different name...

pgroth: What I'm worried about is leaving here with a pseudo-agreement to have an editorial change, and then later someone objects to it
... What Graham wrote seems like a different concept than what we have
... We need an answer from the WG to the question: "Is this construct worth delaying everything else?"

Luc: I think it's different from what's in the document, but essentially, you didn't change the bullets, or did you?

GK: I reordered them
... I said generalEntity: an identifier (supra) for an entity that that appears in bundle (b);
... whereas you said: generalEntity: an identifier (supra) of the entity that is being mentioned.
... and: specificEntity: an identifier (infra) of the entity that is a mention of the general entity (supra);
... instead of: specificEntity: an identifier (infra) of an entity that is a specialization of (supra);
... I couldn't understand the original description, but mine is what I made from it after discussion

Luc: what about incompatibility with RDF semantics?

GK: that was part of the basis of my concern, but not the essence
... I'm checking whether I can make lighter changes with the same effect

<lebot> prov-o's definition: "prov:mentionOf is a special type of prov:specializationOf whose subject presents as an aspect a particular prov:Bundle in which its more general Entity was described (prov:asInBundle is used to cite the Bundle in which the generalization was mentioned)."

tlebo: When we were comparing the bullets, I was thinking it would make sense to keep the current DM definition for bundle and specific entity, but use Graham's general entity

<lebot> generalEntity: an identifier (supra) for an entity that that appears in bundle (b);

tlebo: If we get rid of this word "mentioned", then we can avoid some confusion

<lebot> generalEntity: an identifier (supra) of the entity that is being mentioned.

<lebot> bundle: an identifier (b) of a bundle that contains a description of supra and further constitutes one additional aspect presented by infra.

<lebot> ^^ wipe that :-)

<lebot> specificEntity: an identifier (infra) of the entity that is a mention of the general entity (supra);

<lebot> generalEntity: an identifier (supra) for an entity that that appears in bundle (b);

<lebot> bundle: an identifier (b) of a bundle that contains a description of supra and further constitutes one additional aspect presented by infra.

Luc: I'm concerned that we're not progressing

ivan: I think the only way to move forward is to drop it from the spec
... It's harsh, but realistic

SamCoppens: This seems to be about interpretation. Can't we just leave the description as such, but explain using Graham's example?

ivan: We are at the last minute

pgroth: We're not even arguing about a little bit of text. This is a substantial change
... The goal of the DM was to have an intuitive, easy to understand model.

<lebot> wasInfluencedBy is confusing with wasInformed

<Zakim> lebot, you wanted to ask isn't this what FAQs are for? That's how we addressed the issues earlier today.

pgroth: Now, we agree on the structure, but not on the definition it seems

tlebo: isn't this what FAQs are for? That's how we addressed the issues earlier today.

pgroth: But the commenter from earlier today wasn't a WG member, that had the chance to discuss with us for a long time
... If it's not clear for Graham, how can we expect outsiders to get it?

Luc: yesterday, it seemed to me like there was no support for the construct. But this morning it seemed there was.
... But now we have to move to CR.

GK: I will back down from making a formal objection, after discussing it today

Luc: Still, at previous meetings, we agreed that if there's no consensus, we would drop it.
... I say we just vote

smiles: I wanted to ask: what is the negative consequence of it being removed?

pgroth: You can't use it
... we lose some interoperability

<Curt> we have a 6pm res

jcheney: It might be good to state the pros and cons
... pro: clear use case
... con: it's been controversial

scribe: pro of removing: covering our euphimisms

<smiles> A note has the advantage that if a better way is found later, the DM would still stand complete without the note

tomDN: still in favor of creating a note. seems like the same amount of time, but without delaying CR

hook: Could we do something less strong than that?

Luc: like a wiki

ivan: You could take what's there, and put it into an informative appendix
... as a guideline.
... But it wouldn't be in the standard ontology

<Curt> similarly, we would have to leave it out of the XML schema

pgroth: My worry with that is that it's confusing.
... CR speaks with a clear voice
... An informative appendix does not

Hook: So do we provide ambiguous guidance or no guidance at all?

pgroth: Either crystal clear or not at all

ivan: Any member can do a member submission, but that's really the weakest form

<pgroth> propose: Keep mentionOf as part of PROV as is and not at risk

<smiles> -1

<lebot> +1

<Curt> +1

<jcheney> 0

<hook> +1

<GK> -1

<zednik> +1 (RPI)

<SamCoppens> 0

<pgroth> resolved: mentionOf is removed from PROV rec track documents

<Curt> as a note, what would be the effect on the OWL or XSD schema?

-q

<smiles> @pgroth yes, exactly

Luc: Simon, Graham, would you object to mention as is in a note?

Graham: no, I'd go -0 or support it

smiles: no, probably 0

ivan: so timetable for this hypothetical note?

pgroth: not together with CR. It's a "new"note.
... Who would do this?

Luc: As editors. we should take out the text from the recs, and put it into a document
... I'll take on this

<pgroth> ACTION: luc create a mention of document [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action15]

<trackbot> Created ACTION-135 - Create a mention of document [on Luc Moreau - due 2012-11-16].

Hook: This will be a new note for the DM, but how far deep would the note go regarding the other documents?

Luc: A single, comprehensive document

ivan: What about the 2 extra terms in owl. Which namespace would that be?

<lebot> @hook, the "put it all together" approach is what we agreed to do for dictionary

Luc: same for XML

s/same/same question for

<lebot> didn't we agree that the dictionary term URIs were "reserved" in our namespace?

pgroth: same solution as with the other notes

ivan: has to be made clear that these are not standard properties

pgroth: We have prov-aq.owl, prov-dc.owl, etc.

<hook> @lebot, thanks for clarifying.

pgroth: Eventually, we'll create a "super" owl file including everything, with clear commenting what is standard and what not

ivan: So it'll all be in the same namespace. And I am happy with that

Vote for CR

<jcheney> should we formally close 475??

<pgroth> close ISSUE-475

<trackbot> ISSUE-475 Request to drop "mention" and related elements closed

<pgroth> ACTION: Luc to update public response on mention [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action16]

<trackbot> Created ACTION-136 - Update public response on mention [on Luc Moreau - due 2012-11-16].

<pgroth> proposed: prov-dm, prov-o, prov-constraints, prov-n to be submitted as candidate recommendations as soon as all editorial actions are completed

<jcheney> what is the record for number of issues?

<ivan> +1

<Curt> +1

<smiles> +1

<jcheney> +1

<GK> +1

<hook> +1

<SamCoppens> +1

<lebot> +1

<zednik> +1 (RPI)

<pgroth> +1 (VUA)

<Luc> +1 (Southampton)

<pgroth> accepted: prov-dm, prov-o, prov-constraints, prov-n to be submitted as candidate recommendations as soon as all editorial actions are completed

<smiles> Sorry, got to go now. Talk to you tomorrow

PROV-AQ

<GK> https://dvcs.w3.org/hg/prov/raw-file/12cd1aaa575a/paq/prov-aq.html

Luc: It would be good to hear from the editors and set a time for a next release

GK: Until about a week ago there was no progress.
... In the last week I started going through the issue list
... 25 are pending review
... There were 2 issues I'd like some feedback on
... One is link relations or full URIs

tlebo: If there's something I can edit in the document, I could settle my raised issues

GK: Just let me know what the changes are

ivan: rel="provenance" is something that isn't defined by HTML yet
... if you use full URIs, you don't have that problem

pgroth: can you use those in the header of an HTTP request?

ivan: not sure

GK: I think it might work

ivan: Another option is RDFa
... prov:provenance
... Might be good to talk to the Linked Data Profile WG
... I am not familiar with all the details of their spec, but it makes sense to try and comply with their method
... Making it clear that we arent talking about a REC

tlebo: think the proposed change (to put a full URI or prov: prefix in link/@rel) would actually fix the issue that I ran into in March when trying to use AQ in PROV-O HTML.

GK: So we basically agree to push ahead with URIs
... Paul raised an issue about introducing roles of consumer and publisher
... I've taken that on board in the discovery section, so you may want to review.

pgroth: Locating provenance information section?

GK: yes
... We are also dropping the reference to POWDER

pgroth: Do we still want best practice in this document?

ivan: This also might be interesting to discuss with the LDP WG

Luc: Do want a discussion on bundle identifiers? And how we access their content?
... When are we aiming for the next release?

GK: last time I checked, by the end of this month
... at least with the outstanding issues resolved and ready for another round of review

Luc: So the end of the year would be feasible?
... And do we synchronize with the family of specs?

ivan: Absolutely

pgroth: I would like an implementation of AQ
... using the example corpus of provenance
... Also, the document should be cleaner
... (e.g. best practices inside the document)
... smaller would also be good
... We should aim for a release cycle by the end of the year

Luc: As we did with the DM, we can release an internal draft for review of specific people

GK: I'll give it a shot
... Question for the group: What do we do with the issues that have been there for a long time?

Luc: We should send out reminders

http://www.w3.org/2011/prov/track/products/5

pgroth: I'll set a date for all the pending reviews

GK: sounds good

<Luc> ACTION: pgroth to organize closure of issues closed pending review [recorded in http://www.w3.org/2012/11/09-prov-minutes.html#action17]

<trackbot> Created ACTION-137 - Organize closure of issues closed pending review [on Paul Groth - due 2012-11-16].

Luc: We're trying to close the ones created before summer, specifically
... Anything specific (technical) that you'd like to discuss now?

GK: Not really

pgroth: I just need to respond to your responses

Luc: Do we want to say something about dereferencing bundle identifiers to obtain the content of a bundle?
... Currently, we don't have a mechanism for that

ivan: Intuitively, I'd say you GET a set of provenance statements
... in some serialization
... depending on content negotation
... (RDF or XML)

<Curt> and PROV-JSON!

jcheney: Naïvely, it seems that PROV-N and PROV-XML define what a PROV document is, and that has a name/identifier
... Are we saying that the URIs of the bundles in that document should be dereferencable?

lebot: I propose an alternative HTTP response: at least one triple would come back, saying that the type is prov:bundle

@lebot: (is that about right, Tim?)

<pgroth> that's actually how we do it the paq

Luc: What if the bundle name is not a URL, so you can't dereference it
... We may have UUIDs...

<lebot> I think s/UUID/hash(graph)/ helps phrase the discussion better.

s/UUID/hash(graph)

<lebot> @TomDN No, @luc means UUID.

<lebot> VOID and DCAT handle this distinction with void:dataDump and dcat:distribution [ dcat:accessURL ]

ivan: Coming back to James's question. If we're talking about an ID, do we mean a document or a bundle?
... The file containing the bundles is conceptually different from the bundles
... I'd like to get the bundle in 1 place

<lebot> VOID and DCAT handle this distinction with ?bundle void:dataDump <THE-PROV-ASSERTIONS> and ?bundle dcat:distribution [ dcat:accessURL <THE-PROV-ASSERTIONS> ] .

pgroth: another way to put it is:How do you retrieve the description of an entity?

<lebot> solve the problem for Entity, you've solved the problem for Bundle.

pgroth: It might be out of scope, but we have to look into that
... "Given the identifier of an entity, how do we get the provenance for that? "

<lebot> This sounds more difficult and less finished than "mention"...

ivan: My advice is to sit down with other WGs that specialize in that

<lebot> (but, not a CR...)

GK: there are many things we /could/ specify. But we should focus on the simple stuff first
... So we start to sketch our own thoughts on the matter, and then go to other WGs

<Zakim> GK, you wanted to say there are many things we *could* specify, but there'a a question of how much we *should* specify - we want to guide developers to easy, simple options where

pgroth: To me, we should go and look what LDP does
... Because, in a linked data context, all that stuff is already defined
... Do we want interoperability in this space?
... The linked data community is trying to tackle that, we don't have the manpower for it
... I want to focus on "Is the way we do it, the best, simplest, correct way to do it?"

Luc: conclusion: this issue is out of scope?

GK: We should just be careful about which route we go down on

Luc: So the editors will come up with a lightweight approach

pgroth: I think the best practice should be separate
... That way the document becomes nice and small, and very clear
... and easy to implement
... and then all the bundle/SPARQL stuff separate

<Luc> proposed: PAQ editors to provide a light weight answer to ISSUE-596

<Luc> accepted: PAQ editors to provide a light weight answer to ISSUE-596

<pgroth> trackbot, end telcon

- DRAFT -

F2F Meeting, Stata Center

09 Nov 2012

Attendees

Contents

Admin

Where we are at

Candidate Recommendation for prov-dm

prov-n

PROV-O

PROV-Constraints

Mention

mention of and CR

PROV-XML

prov-dictionary

prov-sem

Mention

Vote for CR

PROV-AQ

Summary of Action Items

Scribe.perl diagnostic output