Provenance Working Group Teleconference -- 22 Mar 2012

<trackbot> Date: 22 March 2012

<pgroth> Scribe: Curt Tilmes

<pgroth> anybody, having problems with Zakim?

<pgroth> curt, thanks for scribing

<pgroth> it's all set up

<pgroth> anybody having problems with Zakim?

<MacTed> I'm not familiar enough with Hg....

<MacTed> can someone tweak the opposite-of-generated example from "prov:consumedBy" to "prov:wasConsumedBy" to more properly parallel "prov:wasGeneratedBy"?

<pgroth> having trouble joining

<MacTed> (it's either add "was" to the consumption, or remove it from the generation...)

<Luc> trackbot, start telcon

<trackbot> Meeting: Provenance Working Group Teleconference

<trackbot> Date: 22 March 2012

<MacTed> Luc has a somewhat bad connection... fuzzes a lot on vocals

Admin

<pgroth> Minutes of the March 15 2012 Telecon http://www.w3.org/2011/prov/meeting/2012-03-15

<dgarijo> +1

<Paolo> +1

<mike_> +1

<zednik> +1

<satya> +1

<GK> +1

<pgroth> Approved: Minutes of the March 15 2012 Telecon

pgroth: time change reminder -- return next week to normal times
... please sign up scribes
... open actions - paul has drafted PAQ review
... graham still needs to review that

<pgroth> +q?

pgroth: he just got it -- he will address this week or next

<GK> I'm tied up with project work this week, but should havew time to review next week.

luc: is the plan to release the PAQ synced with the others?

pgroth: depends -- it may be difficult to do that, graham may be able to get things and we might be able to do it
... we'll at least try to get a synced draft release

<pgroth> ack ?

<GK> (lost my sound briefly then)

<pgroth> https://docs.google.com/spreadsheet/lv?key=0An15kLxkaMA3dFVCWm9aREZFemNOYjlGQjdPRkdFZXc

pgroth: review updated time schedule from google doc
... synced release of docs in mid-april
... drafts available to the group by the end of next week for internal review
... a lot of work has happened, we need the synced release to show progress

Prov-dm

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html

luc: much work in the last week, revised editors draft exists, not quite ready for review

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#data-model-components

luc: structure revised

<Paolo> you'll need to scroll down as usual

luc: section 4 beginning, has a picture of the 6 components
... for each component, we have a short overview and UML diagram depicting the relationships between components

<Luc> : http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#section-example-a

luc: things are taking shape, the essence we wanted to incorporate is there

<tlebo> colors and tetris. All right!

luc: section 3.1 has a revised picture taking into account feedback

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-n.html

luc: PROV-N updated editors draft

<GK> (This linking problem is a respec anomaly -- once page is loaded, clicking on section link works.)

luc: definition of production, some conflicts resolved, taking shape now

<Paolo> @tlebo I like to think it's a tangram :-)

luc: still working on collections -- more expected in the coming week

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#glossary-alternate

<Luc> http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#glossary-specialization

luc: solicit contributions for short english definitions for some concepts
... please email suggestions to luc/paolo and they will incorporate

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

<GK> Do we really need alternates/specialization in part 1?

<Luc> Communication

<Luc> Start By Activity (wasStartedByActivity) http://dvcs.w3.org/hg/prov/raw-file/default/model/prov-dm.html#term-wasStartedByActivity

luc: wasInformedBy relation unclear -- suggesting "communication", read that section of the document

<dgarijo> what about dependency?

luc: next steps to get to release, still going through some feedback

paolo: constraints need some work
... how hard is the deadline next week? many things have to happen in the next week...

luc: it is challenging to get 3 documents all together at once. section 2 needs some work
... help on definitions would be welcome

paolo: part1, 3 are the priority

pgroth: review other documents status, then discuss timeline
... still need to line up reviewers for various documents
... that will help inform timeline

<pgroth> Prov-dm Reviewers

<tlebo> +1

<khalidbelhajjame> +1

<satya> +1

pgroth: need reviewers for each of three documents

<jun> +1

<Luc> @satya: documents are not ready!!!!!!

<pgroth> Prov-dm-constraints Reviewers

<jcheney> +1

<tlebo> +1

<GK> I'll plan to review "constraints", provided it's available before April

<pgroth> Prov-n Reviewers

<khalidbelhajjame> +1

<smiles> +1

<tlebo> +1

Prov-o

tlebo: reorganized HTML document, emphasizing difference between simple/extended

<tlebo> http://www.w3.org/2011/prov/wiki/Meetings:Telecon2012.03.22#PROV-O

<scribe> ... new design in progress, met monday and assigned some writing sections

UNKNOWN_SPEAKER: getting together early next week, draft available next friday

<pgroth> Reviewers Prov-o

<Luc> +1

<pgroth> +1

<smiles> +1

<GK> I think the most important reviews of PROV-O will come from implementers like Stian

luc: for next week, editors of the documents should identify specific questions for reviewers to address

<khalidbelhajjame> @Graham, Stian is part of the provo team :-)

pgroth: this is a review to determine public release, should they go public?

Primer

<GK> @Khalid - Even better!

luc: that is one question, there may be others

smiles: still revising, determine if other docs are stable to incorporate any change back into the primer

<GK> @Curt ^^ That was Simon, I think

smiles: looking for examples to put in

tlebo: use the big example collection online

smiles: examples organized by domain, not by concept -- need to find focused examples for concepts

tlebo: we might try to pick trivial examples for each concept
... look at the descriptions -- it lists the constructs that various examples use

<pgroth> Primer Reviewers

tlebo: might be out of date, let me know if something needs to get updated

<Paolo> +1

<tlebo> coverage page: http://aquarius.tw.rpi.edu/prov-wg/provrdf-owl-coverage shows which constructs are used in which examples (or did, or should...)

<Christine> +1 if you think it would be useful

<Luc> Will review, but may not meet deadline if still editing prov-dm-*

<pgroth> Schedule:

<smiles> @tlebo great, thanks

pgroth: schedule -- editors can you meet the target dates?

<tlebo> @smiles, some of it looks broken, if there's a particualr question you need answered, point it out to me, please.

<GK> Might it be helpful to stagger releases, so reviewers aren't overcommitted?

paolo: two out of three parts, next friday possible, part 2 maybe the next week

luc: yes, I would like to do that

<tlebo> Friday is reasonable target, early the follwoing week is realistic

luc: PROV-DM,PROV-N by friday; constraints early the next week

<GK> It might be difficult for me to get a printed copy for review oif not available by 3-Apr

tlebo: next friday possible, early the next week more realistic for PROV-O

smiles: next friday ok

pgroth: how about Monday the second, but earlier if possible

<smiles> Fine by me

<Luc> review timetable?

pgroth: if we do Monday 2nd, then use 1 week for review

<pgroth> 4/9 for reviews back

<Luc> easter vacations ....

<GK> 1 week for review would be OK for me.

<tlebo> leaves 1.5 weeks to review, no? (was 2 weeks)

<Luc> i can review provo on 4th in time for the 5th telecon

<Paolo> if I am doing just the primer, won't take long. I would like to contribute a new example for collections. So, 2 days

pgroth: the 2 weeks was 1 week to review, 1 week to respond to the review

<tlebo> @paul, got it. 1 week to review and 1 week to reflect feedback.

luc: release/voting had been planned for 12th, probably not realistic with new schedule, esp. with Easter et al.

pgroth: what about 4/19 for formal vote for release

<pgroth> Vote April 19

luc: at the latest

smiles: clarify

pgroth: one of the questions for review -- are there release blockers?
... if there are things that can be addressed later, that's ok

smiles: even non-reviewers need to read enough to determine their vote

<Luc> yes, everybody will be invited to review

<pgroth> Revised schedule:

<pgroth> 4/2 for release to reviewers

<pgroth> and working group

<pgroth> 4/9 reviews in

<pgroth> 4/19 vote by working group

<tlebo> looks fine

<Paolo> works for me

<satya> works for me

<jun> +1

<jcheney> I may need another day or two = the period from 4/2-4/9 is super busy for me

luc: potentially earlier, perhaps 4/12 if possible

<khalidbelhajjame> good for me

<smiles> Yes, possible I think

pgroth: reviewers should focus on question of release

Namespace Unification

<pgroth> http://lists.w3.org/Archives/Public/public-prov-wg/2012Mar/0386.html

<pgroth> http://www.w3.org/ns/prov#

pgroth: put out proposal^
... that would point to landing page with glossary of all the terms
... content negotiation would return OWL/TTL/XSD as requested

<Zakim> GK, you wanted to ask Do tools that try to read ontologies generally generate accept headers?

<MacTed> voice is there, but it's not OK.... lots of breakup. some is comprehensible, most is not.

gk: <breaking up> will the tools be able to retrieve the right thing?

<tlebo> @gk, AFIAK, yes. Tools do request RDF. If they don't, they should.

<Luc> http://www.w3.org/2001/XMLSchema# in rdf

<Luc> http://www.w3.org/2001/XMLSchema in xml.

luc: I'm fine with pgroth proposal, but we also had a namespace for the XML schema, is the hash compatible?

<GK> AFAIK, The choice of XML namespace for XSD preceded the use of namespaces for RDF, and the issue of concatanating namespace+local to forkm a URI didn't arise.

pgroth: if you use RDF, you use the hash, with XML, you don't [??] you still dereference to the same place

tlebo: you have to remove the fragment when you request the URI with HTTP

<pgroth> sandro are you there?

tlebo: others use the hash the way we are proposing

<GK> (I('m looking in XML namespace spec at moment)

<zednik> '#' is for client side processing, yes?

<pgroth> : http://www.w3.org/2001/XMLSchema#

luc: not sure -- it is frustrating to have to use two different namespaces

<pgroth> : http://www.w3.org/2001/XMLSchema##

<sandro> (I am here, pgroth and trying to page in the details)

pgroth: XML usage would double the hash

<GK> [Definition: An XML namespace is identified by a URI reference http://www.w3.org/TR/REC-xml-names/#URIRef; element and attribute names may be placed in an XML namespace using the mechanisms described in this specification. ]

<GK> That's from XML namespace spec

<sandro> http://www.w3.org/2007/rif

<GK> That allows a '#' in the namespace URI for XML namespaces

<GK> http://www.w3.org/TR/REC-xml-names/#concepts

sandro: RIF used the hash within the XML usage, didn't get pushback

<Zakim> tlebo, you wanted to state that # is part of the fragid, and is taken off before HTTP request. Also, all w3 namespaces uses the #.

sandro: the hash is very normal within RDF community

macted: hash is problematic, esp. since it is really a client side thing

<GK> Tools that send '#' are BROKEN

macted: a small number of tools send it to the server

<pgroth> can you not talk into your mike

pgroth: how can we best resolve this?
... this plan sounds good, but how can we determine if it is correct?

sandro: I'll consult with XML experts, schema workgroup, etc.

pgroth: semantic web activity group?

<GK> What we could do for now is choose a '#' URI and document our rationale that this is OK in XML (see above), and ask for explcit review in PWD.

sandro: this is really an XML question
... could also ask semantic web coordination group

<tlebo> http://www.w3.org/ns/prov#used refers to #used at http://www.w3.org/ns/prov -- isn't this the same as what luc's xml schema is referencing?

<Zakim> GK, you wanted to note that my reading of spec says '#' OK in XML namespace ... see notes above

sandro: will draft an email requesting input from others

tlebo: namespace shouldn't have the hash

sandro: do any tools break?

<GK> RDF really likes the namespace to end with '#' or '/' - so that the parts can be teased apart later.

luc: esp. with RDFa, you are using RDF and XML together, namespace declarations just define 'prov'

<sandro> <prov xmlns="http://www.w3.org/ns/prov#">...

<MacTed> are there any known examples where NOT having the # in the namespace leads to problem?

<MacTed> we've come up with various ways having it *MIGHT* cause trouble...

<MacTed> so why do we *want* it to be part of the namespace?

<MacTed> what's the argument to include it?

<GK> RDF/XML commonly uses XML namespace decls with '#' at end of URI

<Luc> http://www.jenitennison.com/blog/node/49

luc: see blog^ she talks about this problem
... maybe she could help us

<GK> Yes, that;s the issue - RDF concatenates, XML doesn't assume the pair are used to create a new URI

pgroth: let's consult with all these groups/people to figure out the right common namespace

<GK> I think a common namespace is nice, but if we can't it's not a total disaster IMO

sandro: will review the blog and include jeni as well

<GK> http://www.jenitennison.com/blog/node/49 - this may be Jeni's blog post

<GK> Dated 2007, "Things that make me scream ..."

<pgroth> trackbot, end telcon

- DRAFT -

Provenance Working Group Teleconference

22 Mar 2012

Attendees

Contents

Admin

Prov-dm

Prov-o

Primer

Namespace Unification

Summary of Action Items

Scribe.perl diagnostic output