Provenance Working Group Teleconference -- 17 Jan 2013

<trackbot> Date: 17 January 2013

<dgarijo> Luc, if nobody volunteers, I can be the scribe

<Luc> it's very kind daniel, first, let's see if we can find somebody else,

<Paolo> HI I have done it quite recently but can do it again

<Paolo> if needed

<Luc> scribe: tlebo

(I think this IRC client has dropped me before, so I might need to hand off to someone mid-stream)

<pgroth> ivan has the loudest keyboard ever

<ivan> :-(

:-)

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

<dgarijo> 0 (I wasn't there)

proposed: approve minutes

Admin

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

<ivan> +1

0 (was not here)

<Curt> 0 (not present)

<Paolo> 0 (missed it)

<dgarijo> 0

<smiles> +1

<Dong> +1

<Luc> resolved: last week's minutes

luc: action on Tim for cross referencing.

tim: I'll look at the cross reference.

luc: action on himself...
... action on Stephan for namespaces.

zednik: have looked at identifiers

pgroth: I've done all actions on me.

Implementations

luc: paul went through surveys recently.

<pgroth> http://lists.w3.org/Archives/Public/public-prov-wg/2013Jan/0073.html

<dgarijo> I've added all the remaining implementations that I had added in the wiki. Some of them today

pgroth: announced it via email
... reports from 9 organizations
... southampton has 7 implementations.
... provo is good track; all constructs supported by 2 sep implementations.
... provo needs a pair of implmeentations to exchange.
... it's easy to hit that min by having SH and Kings to exchange prov.
... for prov-n it is less satisfying. only 2 orgs: SH and weblab.
... tom and sam have prov-n implementations.
... but do weblab actually support it?
... they say openRDF sesame as the library (implies prov-o)

<Paolo> I am going to add implementation with prov-n support

pgroth: need to rely on tom/sam implementation

<Paolo> (applications)

pgroth: constraints: luc's implementation, but we need one more impl.

<pgroth> it's a good thing!

luc: # implementations on southampton, some on different languages.
... we can demonstrate language independence if not org independence.
... TomDN implementing? isn't he using the provtoolbox?

TomDN: not really using the toolbox.

<pgroth> ace paolo

Paolo: Tom's note, if I produce prov-n w/o provtoolbox then it counts as a new implementation.

luc: ideally, we want full independence.
... but we made the case that prov-n is aimed at human consumption.
... so no need for interoperability.

pgroth: for prov-n, does TomDN consume prov info as prov-n?

<Luc> prov-dm contains examples of prov-dm, consumable by prov toolbox. So it's a pair!

<Luc> prov-dm contains examples of PROV-N, consumable by prov toolbox. So it's a pair!

TomDN: I read it == I evaluate. It generates it == write.

<GK> A precedent we might look at is the "human" syntax(es) for OWL - what do they do?

<Luc> @TomDN, can you validate your prov-n with my validator?

TomDN: I can produce a file from my tool as prov-n.

<Luc> @TomDN, can you paste your prov-n into http://openprovenance.org/prov/validator.html

pgroth: we want to give TomDN's prov-n to luc's validator.
... then we have a pair.

<Luc> or http://openprovenance.org/prov/translator.html

pgroth: tom doens't use all of the constructs.

TomDN: no.

pgroth: extend to use other constructs?

TomDN: not easy.

Luc: prov-dm doc is an example of prov-n generation. so all statements can be validated by toolbox.
... thus an interop pair

pgroth: we need to add it to the quesionnaire

<TomDN> (I will provide some rationale on my questionaire answers to explain the read/write answers

pgroth: we should do it as the WG

<GK> Losrt my client

pgroth: we seem to have coverage on prov-n (with these steps)

<GK> Sorry lost connection

Luc: please submit the implementation reports.

GK: prov-n and interop -- find out what OWL WG did with their functional syntax.
... it's similar role.

<ivan> unmute ivan

<Zakim> GK, you wanted to suggest looking in to how http://www.w3.org/TR/2012/REC-owl2-syntax-20121211/ was handled

<pgroth> i think we have a solution now for prov-n

ivan: i don't remember how function syntax was handled. it was just a specification language - not a cand rec requiement to parse it.
... for OWL systems, they needed to exchange RDF/XML. the ONLY one.

<GK> That was my guess - I think PROV-N might be presented similarly?

<pgroth> Tom's implementation + validator and prov-dm docs + validator

<Zakim> pgroth, you wanted to say we need a pair for prov-o

pgroth: provo coverage is fine, we need some pairs.
... could simon and luc make that pair?

Luc: the service is available online. anyone can paste them in to validator or translator.
... they can just report it.

<smiles> Luc - Can you give the URL again?

<stain> Zakim: ??P44 is me

pgroth: a pair isn't one directional?
... simon, dump out provo and validate it?

smiles: yes, if it takes turtle

<Luc> http://openprovenance.org/prov/validator.html

Luc: anybody can do that.

<pgroth> constraints

<GK> I think producer -> consumer (one way) is OK - it shoews two developers read spec and had some common understanding

pgroth: constraints. luc did them, pgroth is slowing working them.
... obviously good to have more than 2
... or even if other parts of implementations.
... 47 unit tests away from full coverage.
... shoiuld be able to do it.

+1 :-)

luc: Paolo was working on partial constriants

<TomDN> Does your implementation need to be completely finished by Jan. 31st?

Paolo: jun and I havent' worked on it.

<pgroth> @TomDN it's about the report really

Luc: when? by end of month?

Paolo: not by then.

<pgroth> +q to ask about stardog

<TomDN> @pgroth, thanks, that's what i thought

pgroth: we knwo there are implementations out there than what is reported.
... we should all sign up to report implementations.
... e.g. stardog is doing it.
... approach individually.

<Zakim> pgroth, you wanted to ask about stardog

(I tried to poke the LInkedTV via twitter yesterday -- it's not coming out till Feb)

<pgroth> but there's also dbpedia, and qudt, etc

Response to public comments

<pgroth> simon can you mute

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

<GK> (There's also the possibility of submitting "third party" reports for public implemenbtations)

<pgroth> @GK i didn't know that

<Luc> Proposed: the group endorses the responses to issues 611 and 612

<TomDN> +1

<ivan> +1

<khaidBelhajjame> +1

<satya> +1

<zednik> +1

<stain> +1

<Dong> +q

<TallTed> +1

<GK> @paul - I don't *know* that, but I can't see any reason why now

<SamCoppens> +1

<dgarijo> +0 (I haven't reviewed them yet)

<hook> +1

<smiles> +1 (noting that I sent a suggestion just before this call)

<GK> +0

<Paolo> 0 haven't reviewed in detail

<Luc> Approved: the group endorses the responses to issues 611 and 612

<Curt> +0 (haven't reviewed)

<stain> @TallTed what happened to McTed over New Year? :)

luc: who to respond to reviewers?

<pgroth> sure

Luc: : is it paul?

pgroth: I can do that.
... we had a question about derivation.
... wasQUotedFrom, not derivation.

<dgarijo> I think it was about the name of wasQuotedFrom.

(I think it's the same mix-up that "Tim vs. Stian+Daniel" had.

<stain> it was about the directionality - as I originally complained about it

<stain> he suggested hadQuoteFrom

<dgarijo> yep

<pgroth> ACTION: pgroth to respond to public comments [recorded in http://www.w3.org/2013/01/17-prov-minutes.html#action01]

<trackbot> Created ACTION-162 - Respond to public comments [on Paul Groth - due 2013-01-24].

<Dong> @Luc and Tim: I've just remember that some PROV-O examples might need to be revised in ISSUE-611

luc: wasquotedFrom, can't recall.

<TomDN> http://lists.w3.org/Archives/Public/public-prov-comments/2013Jan/0006.html

<stain> I think we need another WG discussion on email about it

<TallTed> stain - just unifying my nick across a few spaces (Twitter, a couple of IRC nets, etc.) ...

<stain> if the term was confusing in the primer.. then it's confusing all over

ECHO

<pgroth> echo madness

<pgroth> simon

<Zakim> tlebo, you wanted to note that we couldn't converge last time on the naming.

<smiles> @tlebo - Do you have a link to that email conversation?

<stain> but this email was from Chuck Norris - which sounds like a native speaker

smiles: I can dig up the ISSUE.

<stain> MORRIS

<stain> hihi

<smiles> OK

<dgarijo> @stian:lol

prov-dictionary

tlebo: I'll dig through them. thanks for the reminder

<TomDN> no

smiles: https://www.w3.org/2011/prov/track/issues/352 is the wasQuotedFrom naming discussion

<Paolo> deadline is 23rd, right?

<pgroth> fine with me

Luc: reviewers should give feedback by next wed.

<pgroth> paul

<smiles> @tlebo Thanks!

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

prov-aq

GK: I've seen 4 reviews in

pgroth: tim, simon, luc, dong, stian

<pgroth> your back

pgroth: all have reviewed.

<pgroth> shall i go

GK: through 1.5 sets of comments.
... most are editorial.
... hoping to pick out those that are more than editorial.

<pgroth> that's not the main issue

GK: issue on REST interface vs. simple convention for URI to retrieve prov.

pgroth: everyone but stian said doc can go LC
... stian has 8-9 blocking issues.
... we should concentrate on blocking issues from stian.

<GK> Ah, I hadn't yet looked at those blocking issues from Stian

stain: my issues: it was heavy.
... as a draft, fine. but not as final technical.

<stain> @GK sorry about that..

Luc: i'm fine wiht doc released as next WD.
... I felt that feedback required changes, and so not last call.
... we don't have notion of last call in Note.
... we're abusing that name.
... as next draft, fine.

pgroth: we should distinguish between release and not wanting to do any more to it.
... address all comments, release a WD at a minimum. OR when we go PRec with other docs.
... but can we get another iteration?
... before PR

<GK> @paul +1 review/revise as much as possible

pgroth: we need two more itnerations on PAQ

Luc: we should be pragmatic about Notes and making it too perfect.

<GK> If the issues are substantive, I don't think they should be "offline"

<pgroth> @gk i was talking about scheduling

<GK> @paul Ah, OK.

Luc: do we vote? or do work and editors cycle?

pgroth: next week or following for WD.

<TomDN> (can be synced with dictionary then)

pgroth: take off "final" terminology"

prov-xml

<GK> Next week is likely busy for me. I'll try to complete my pass through the reviews today.

luc: namespace, schema management, etc.

zednik: some feedback on original note. extended types was confusing with prov:type.
... made native XML type for those.
... identifiers: work natively for XML
... id and idref

id uses xsi:id

scribe: base type of idRef -- xml tooling is familiar.
... cannot start with numbers, other constraints.
... it's native but doesn't work with our examples or URIs.
... alternative: anyURI or QNames

<pgroth> @gk see where you can get and see what needs to be debated

scribe: not sure which would be better.
... also xlinks and xpointers.

<GK> @paul - that's my plan - I'm making notes as I go. I'll email you a copy when done.

<pgroth> @gk awesome

scribe: xpointers and xlinks might let us verify references existing.
... the group needs to read up on xlinks/xpointers.

ivan: how widely is xlink implemented?
... do tools really do it? xlink is an unlucky standard.
... might not be worth adopting, could be more harm than good.

hook: xlinks ISO community uses them to reference external XML traces.
... in bundles, can reference across bundles.
... but good point on how much it's used.
... we still need to look into xlinks.

Luc: have your changes surfaced to the WG?
... some may want to get insight into your changes.
... (are you using issue tracker?)

zednik: announced, but did not tag into ISSUE.
... last modeling was valid.

Luc: namespace management issue
... the namespace HTML is waiting on it.

zednik: seems like we're misusing xml namesapces.
... the extension should have a different namespace.
... including dictionary in new namespace
... stian proposed an organization, but not ideal.
... the xml schema should have different namespaces.

<TomDN> (didn't we vote on this a few months ago?)

pgroth: prov dictionary etc being in different docs doesn't mean it's not in the same thing.
... (they are in same group, it's just broken up to aid understanding).

<Luc> zakim. who is noisy?

pgroth: from practice, developers do not like separate namespaces.
... PROV namespace will included in RDFa 1.1 automatically.

Luc: is nice to have one ns
... a prov-xml dictionary namespace?

<TomDN> then we need a prov-links namespace as well

<GK> IMO, if any of the XML namespaces are different from the corresponding RDF, then they should *all* be different

<GK> .. including the "core" namespace

<Curt> just do 2, core, and core + all extensions

<stain> exactly.. we do a single schema with only core, and one with core-everything

stain: oen for core prov xml, one that is prov-everything xml schema. If you don't like either of those, then you're on your own.

<Zakim> tlebo, you wanted to ask zednik where the motivation for different namespaces comes from

<Curt> I'm ok with one schema with everything

<pgroth> i like that

Luc: : single schema file with all terms that use one namespace, THEN half-way house of a single schema for just the core Rec terms.

<stain> I think this is mainly an artifact of XML Schema being very strict of linking schemas and namespaces - this is not a big deal in other ways to express XML schemas like Relax NG

<Zakim> pgroth, you wanted to say that it is a single namespace

pgroth: everyone should be able to paste the namespace to a browser and get all serializations documented.

<khaidBelhajjame> bye

<GK> Bye

<SamCoppens> bye

<dgarijo> bbye

Luc: I am done scribing?

<Luc> trackbot, end telcon

- DRAFT -

Provenance Working Group Teleconference

17 Jan 2013

Attendees

Contents