Meeting Record

There are some format problems with the chatlog. Please correct them and reload this page. They are labeled on this page in a red box, like this message.

It may be helpful to

Scribe problem: the name 'Kazuyuki' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'ericP' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Ralph' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'EricP' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'olaf' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Klaus' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Matt; got it' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'matt' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Zakim IRC Bot: the conference code is 4562 (tel:+1.617.761.6200 tel:+33.4.26.46.79.03 tel:+44.203.318.0479), matt ←

Scribe problem: the name 'Ivan' does not match any of the 0 active names. Either change the name used, or request the list of names be altered.Active names:

Zakim IRC Bot: Not knowing who is chairing or who scribed recently, I propose Rigo ←

Scribe problem: the name 'Ivan' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'pgroth' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Yolanda' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Yolanda: goal is to get feedback on architectural issues in handling provenance ←

Scribe problem: the name 'tlr' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Zakim IRC Bot: the conference code is 4562 (tel:+1.617.761.6200 tel:+33.4.26.46.79.03 tel:+44.203.318.0479), tlr ←

Scribe problem: the name 'Steven' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

[I think an example, e.g. a museum collection reference, would help here explaining the importance of documenting provenance] ←

[modern-day example; provenance documentation has been around for much longer than blogging ;) ] ←

Unknown Yolanda: two other scenarios; disease propagation and business processes ←

Scribe problem: the name 'Shawn' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'TimBL' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Luc' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'ivan' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... at the bottom is a representation of the artifact the user has requested ←

Scribe problem: the name 'Doug_Schepers' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Zakim IRC Bot: On the phone I see Keio, Coralie, Jeanne, Ivan, francois (muted), Klaus, Dsr, Rigo, Bert, Steven, Thomas, MIT531, Ted (muted), EricP, olaf, Yolanda, Luc, Sandro, Matt (muted), ←

Zakim IRC Bot: On IRC I see Steven, pgroth, Bert, plh, dsr, betehess, jeanne, Klaus, koalie, francois, ericP, shawn, RRSAgent, matt, Zakim, Ian, YolandaGil, shepazu, jeff, rigo, kennyluck, Ralph, ←

Unknown Luc: limitations on size of HTTP headers make them a less than ideal approach ←

Unknown Luc: imagine you received this news article without embedded provenance; how would you find provenance? ←

Scribe problem: the name 'jeff' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'dsr' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Sandro' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Sandro: this presentation clarified a lot of issues on provenance in my mind ←

... I have a concern: when talking about the protocol, it sounded like the earlier point that provenance is a kind of metadata was ignored ←

Unknown dsr: q+ to ask about provenance and domain names which change ownership over a period of years ←

Scribe problem: the name 'timbl' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... any HTTP header that is added should be a metadata header, not just a provenance header ←

... also, if you do provenance-by-value I don't see how to do provenance on provenance ←

Scribe problem: the name 'PaulG' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown PaulG: I think you can have provenance on provenance but you end up stopping somewhere ←

... generally you'd stop when you find provenance asserted by an individual or by an entity ←

Unknown Sandro: the fact that something is asserted by an individual is itself asserted by an individual in every case ←

Scribe problem: the name 'Tim' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'jose' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown dsr: That resolves into an issue of trust relationships - who do you trust to make provenance statements ←

Unknown Luc: we could use a general metadata field but there may be some issues specific to provenance ←

Scribe problem: the name 'Jeff' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... longer-term, once provenance is much better established, what will provenance look like? ←

... a million provenance services each doing a small piece or 2 or 3 big ones that do everything? ←

Unknown timbl: q+ to say (a) good presentation, thank you -- good framework about metadata in general. Agree with what Sandro said, xcept the worry about met^n data, as you can ground a metadata document in a self-describin things. I feel the provenance is a *part* of the web of data -- having the proenance "by value" can never carr th whol eweb which basically you need -- infor about individuals and operations and events and licences anc contracts and opearyions and tools. ←

Scribe problem: the name 'rigo' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown rigo: I wonder about country borders as provenance. All governments dream of re-instating borders on the borderless Internet. Is this the ultimate tool for the great chinese firewall? ←

Scribe problem: the name 'Paul' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... e.g. the NYT might provide a set of provenance then point to another site as a source ←

... each site would provide what they know rather than aggregating in a central service ←

... could do centralization but from an architectural perspective, each Web site provides what provenance they can ←

Unknown Jeff: if each one does the provenance it can, are there higher-level protocols between them that are required? ←

Scribe problem: the name 'sandro' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown sandro: (I think if we didn't have RDF already, a reasonable design for provenance would require designing it.) ←

Unknown Luc: we envision a kind of graph that is distributed across many servers ←

... the protocol used to retrieve provenance is the same as the Web protocol ←

Unknown timbl: Have a general set of ontologies plus a generic way of saying what sort of information is available elsewhere, like foaf:made vs foaf:pubs. ←

Zakim IRC Bot: dsr, you wanted to ask about provenance and domain names which change ownership over a period of years ←

Scribe problem: the name 'Dave' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... provenance information involving domain names needs to have a treatment of time ←

... we assume at the start there won't be provenance; then we get to archival Web data ←

Zakim IRC Bot: timbl, you wanted to say (a) good presentation, thank you -- good framework about metadata in general. Agree with what Sandro said, xcept the worry about met^n data, as you can ←

Zakim IRC Bot: ... ground a metadata document in a self-describin things. I feel the provenance is a *part* of the web of data -- having the proenance "by value" can never carr th whol eweb ←

Zakim IRC Bot: ... which basically you need -- infor about individuals and operations and events and licences anc contracts and opearyions and tools. ←

... agree with Sandro's point that an HTTP header should be a general metadata header ←

... a tool and a person are both entities about which you can find out a lot more information ←

Unknown sandro: +1 Tim: you need to follow the links in the provenance out to the whole semantic web. ←

... I do not share Sandro's concern about provenance on provenance on provenance; I think you can ground out eventually ←

Unknown Sandro: I was referring only to the case of provenance-by-value; I'm sure we will not limit ourselves to just that ←

Unknown ericP: q+ to suggests that associating "Existing Provenance Vocabularies" with use cases might guide help people understand the space, e.g. MAIMI for microarrays and technical trust or mail-headers for chain-of-custody forensics ←

Unknown sandro: Tim, my prov-of-prov concern is ONLY if you do by-value. WIthout my-reference this can't work. ←

Scribe problem: the name 'Coralie' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'RS' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown RS: why certain relations are in a graph, a piece of SW and we just never finished the work you are attacking now [ Scribe Assist by Unknown rigo ] ←

Unknown rigo: ...searched for powder and did not see much. Would be interested in how much XG has looked into Powder and how many use cases would be covered by Powder ←

Unknown sandro: I'm thinking the one bit of protocol needed is a way to tag a representation with a (linked data) URI. Tagging the RESOURCE, as rel=meta does is wrong. ←

Unknown timbl: (Tim wonders how many POWDER clients there are -- would like to try ti in Tabulator -- which does provenance handling) ←

Scribe problem: the name 'Olaf' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... we viewed POWDER as only for making statements about resources that have URIs ←

... I think it is difficult to refer to representations of resources using POWDER ←

Zakim IRC Bot: ericP, you wanted to suggests that associating "Existing Provenance Vocabularies" with use cases might guide help people understand the space, e.g. MAIMI for microarrays and ←

Zakim IRC Bot: ... technical trust or mail-headers for chain-of-custody forensics ←

Scribe problem: the name 'Eric' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'francois' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown francois: [POWDER would use the same Link: HTTP header as the "pass by reference" scenario, no?] ←

Scribe problem: the name 'kennyluck' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Eric: on the slide with provenance vocabularies; how much does this map help guide people's understanding of the space? ←

... e.g. microarray analysis at one end of a spectrum and mail headers at another point on the spectrum ←

... have you identified a mapping of specific vocabularies to each use case? ←

... e.g. if you use a term from one vocabulary and want to know if there's a related term in another vocabulary ←

... we thought it would be most useful to help people relate terms between vocabularies ←

Scribe problem: the name 'Yoland' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'YolandaGil' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Sandro: the answer on protocol is just needing a way to identify the representation ←

... with this, everything else is just RDF; no need for new protocols, just vocabularies ←

Unknown Paul: we agree there's probably not a need for very complex protocols ←

Unknown Yolanda: the XG provided feedback to next generation RDF; it's important to be able to identify parts of RDF graphs ←

Unknown Tim: worried about the assertion that you can't use existing machinery ←

... Web pages with advertisement; if the advertising is not important, just use the URI of the page ←

... if your provenance data is a dump of a CMS, each resources would get its own URIs ←

... in POWDER I can't imagine that you'd have one name for, say, a javascript module, and a provenance statement that says "get provenance from this SPARQL server", I would not expect different representations of the metadata at different times ←

Unknown sandro: Okay, that's an interesting design choice, from Tim. Identify representation (which is just a string after all -- odd to give it a URI) -vs- Link mutable resources with immutable resources (with their own URLs). ←

... I'd expect to be able to find all the [metadata] information from the same source ←

... a content-negotiated resource would come with a Location: header that specifies, e.g. a GIF representation vs. a PING representation ←

Unknown Olaf: with resources that change frequently the representation of the RDF graph will change often ←

Unknown Tim: if the database has a journalling system you just append the length of the journal ←

Scribe problem: the name 'Ian' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... probably would just want URIs for versions in the CMS, not for every timestamp ←

Unknown Ralph: I just ewantedt re-emphasize the poin that URIs an only refer to physical things -- they cn refer to conceptial things [ Scribe Assist by Unknown timbl ] ←

Unknown RS: re-emphasize the point: misperception that URIs can only refer to physical things, they can also refer to conceptional things [ Scribe Assist by Unknown rigo ] ←

Unknown Ralph: We have a lot of prorocol alreay to say "I can't return the thing you asked for, but I can return you things about that thing (like a person)". [ Scribe Assist by Unknown timbl ] ←

Unknown rigo: ... can't return you TimBL, but can return lots of information about TimBL, ←

Unknown rigo: ...methods exist that allow to reference multiple representations ←

Unknown Ralph: Anything which you can think about you can give a name to -- so when an artefact changes over time and has many reporesentations, that should be no problem [ Scribe Assist by Unknown timbl ] ←

Unknown rigo: ...machine of web will allow you to exchange info about a thing that you've given a name. Thus time etc is not a fundamental problem as long as it can be addressed by URI ←

Unknown timbl: ... We have a lot of deplyed machinery or dealing with these URIs. ←

... but a lot of people have a great deal of difficulty getting their heads around this ←

... since they have an idea of what http: "usually" means, they think 404 means "broken" ←

... the tricks we've used in the SemWeb such as "#me" on the end are so hacky in my view as to be harmful, since people think they are broken ←

... I know people who think that if there is no ID with the value "me" in the page, something is broken ←

Unknown ericP: +1 to "there's a cost to changing fragment behavoirs across media types" ←

Unknown Ivan: what is the timing for the XG? when might we get an idea of what a WG charter might be? ←

Unknown pgroth: think it would be good for the prov-xg to look again at powder ←

Scribe problem: the name 'Dsr' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Yolanda: the XG was extended by 2 months; we now end on 30 Nov and I do not anticipate a charter before then ←

Scribe problem: the name 'Keio' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Sandro: I'd like to hear what decisions remain to be made for a charter ←

Scribe problem: the name 'Matt' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... but you should know that we have had limited participation from industry ←

... so we're trying to get a sense of what would help get industry involvement in a WG ←

... I know provenance is something they're very interested in. I have a contact ←

... NASA has a lot of interest in managing provenance for the huge amount of data they have ←

... we want a language that can do simple things but it's hard to get agreement on specifics ←

Unknown Yolanda: we've not discussed that -- I can't give you a sense of what the consensus might be, though my personal preference is RDF ←

Unknown Paul: most people would be happy to have a common way to represent and exchange provenance and it would be RDF ←

... the question is whether we have a consensus on where to start and whether we have a starting vocabulary ←

Unknown Sandro: [hope there are not] people in the group who question whether to use RDF? That would waste a year. ←

Unknown Paul: some have said "we want RDF and something else"; perhaps an XML version ←

Scribe problem: the name 'Jeanne' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Unknown Sandro: if you use RDF then you can use SPARQL; otherwise you'd have to reinvent all that ←

... I think the only way to proceed with this work is to mandate RDF but sometimes that is a political lightning rod in some communities ←

... write the charter the same way this presentation was done; assume RDF, SPARQL, etc. ←

... say "the group will define an ontology and the capabilities of a SPARQL endpoint" ←

Unknown Ivan: I scanned the mailing list and saw a lot of existing vocabularies ←

Unknown Tim: W3C is specifically designed to help solve exactly this challenge of reconciling existing work ←

... not charter a group to make a new ontology but to pick and choose terms from existing ontologies ←

... but if FOAF, OWL, etc. already have terms you need then the WG can just select them ←

Scribe problem: the name 'Rigo' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... we don't have a term for a new ontology that collects terms that already exist ←

Unknown Yolanda: should we try now to say "this is our core set of terms" or is it too early? ←

... I'd expect the charter to say 'the resulting set of terms will allow you to specify the following ....' ←

Unknown Tim: we need to understand how ontology development groups should work ←

... could consider terms as being members of a set, with each term having a 'state' ←

Scribe problem: the name 'Bert' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... I was thinking about an abstract model that could be serialized in several ways ←

... can more easily talk about things that could be awkward to express given serializations ←

... if you're doing something very mathematically sophisticated I might worry ←

Unknown Luc: the challenge we saw was the sort of reasoning you might do over the structures ←

Scribe problem: the name 'Time' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

... a lot of people find rules easier, especially for mapping between ontologies ←

Unknown Ivan: agree that this area calls more for a rules-based approach than for OWL ←

Unknown ivan: Thanks a lot to Yolanda, Olaf, Paul, and Luc, it was a great presentation ←

Unknown Tim: by "ontology" I did not necessarily mean "OWL"; I meant just "a set of terms" ←

Scribe problem: the name 'Ted' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Sandro.a' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'MIT531' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Thomas' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'koalie' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'Jose' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Scribe problem: the name 'pgroth?' does not match any of the 3 active names. Either change the name used, or request the list of names be altered.Active names: Zakim IRC Bot Trackbot IRC Bot RRSAgent IRC Bot

Zakim IRC Bot: Attendees were Alexandre, Jeff, Philippe, Ralph, Doug_Schepers, Karen_Myers, Shawn, Matt, Coralie, Kazuyuki, Jeanne, Kenny, Ivan, Ian, francois, Klaus, Mcf, Dsr, Rigo, Bert, shadi, ←

Zakim IRC Bot: ... TimBL, Steven, Thomas, Philipp, Ted, EricP, +49.302.093.aaaa, Yolanda, olaf, Luc, Sandro, pgroth?, Jose ←

Provenance Project Review

Minutes of 28 October 2010

1. Discussion

2. post-Meeting discussion