IRC log of rdf-wg on 2011-09-15

Timestamps are in UTC.

16:01:37 [RRSAgent]
RRSAgent has joined #rdf-wg
16:01:37 [RRSAgent]
logging to
16:01:42 [sandro]
RRSAgent, make logs public
16:01:45 [Zakim]
Zakim has joined #rdf-wg
16:02:01 [sandro]
Meeting: Provenance and RDF Graphs -- Coordination Telecon
16:02:13 [sandro]
zakim, this is rdfg
16:02:17 [Zakim]
ok, sandro; that matches SW_RDFWG(GraphsTF)12:00PM
16:02:52 [sandro]
sandro has changed the topic to: Sept 15 -- Provenance Task Force --
16:02:58 [sandro]
zakim, who is on the call?
16:02:58 [Zakim]
On the phone I see ??P0, Sandro
16:03:16 [kai_]
kai_ has joined #rdf-wg
16:03:31 [Zakim]
16:03:50 [Zakim]
16:03:51 [Zakim]
16:04:03 [sandro]
zakim, ??P2 is Paul_Groth
16:04:03 [Zakim]
+Paul_Groth; got it
16:04:09 [kai_]
zakim, ??P4 is me.
16:04:09 [Zakim]
+kai_; got it
16:04:28 [AZ]
AZ has joined #rdf-wg
16:04:39 [pgroth]
pgroth has joined #rdf-wg
16:04:58 [Zakim]
16:05:06 [AndyS]
zakim, ??P46 is me
16:05:06 [Zakim]
+AndyS; got it
16:05:36 [sandro]
zakim, who is on the call?
16:05:36 [Zakim]
On the phone I see ??P0, Sandro, Paul_Groth, Scott_Bauer, kai_, AndyS
16:06:07 [sandro]
zakim, drop ??P0
16:06:07 [Zakim]
??P0 is being disconnected
16:06:09 [Zakim]
16:06:33 [Zakim]
16:07:14 [Scott_Bauer]
plans to attend: Antoine Zimmermann, Ted Thibodeau, Andy Seaborne, David Wood, Gavin Carothers, Sandro Hawke, Steve Harris ** may attend: Ivan Herman, Scott Bauer, Pierre-Antoine Champin
16:07:22 [gavinc]
zakim, mute me
16:07:23 [Zakim]
gavinc should now be muted
16:07:33 [Scott_Bauer]
ok sorry
16:07:54 [pgroth]
Luc Moreau, pgroth, Kai, Paolo, MacTed
16:08:19 [gavinc]
I don't think Ralph is going to be here, wasn't sure last night
16:08:35 [Zakim]
16:08:35 [AndyS]
Agenda --
16:08:36 [Luc]
Luc has joined #rdf-wg
16:09:29 [AndyS]
Material --
16:09:57 [Zakim]
16:10:02 [sandro]
Luc Moreau, Paul Groth, Kai Eckert, Paolo Missier, Ted Thibodeau
16:10:44 [sandro]
16:10:48 [Zakim]
16:11:01 [Paolo]
zakim, ??P52 is me
16:11:01 [Zakim]
+Paolo; got it
16:11:04 [sandro]
Guest; Luc Moreau
16:11:09 [sandro]
Guest: Paul Groth
16:11:19 [sandro]
guest: Kai Eckert
16:11:28 [sandro]
guest: Paolo Missier
16:11:37 [sandro]
guest: Ted (MacTed) Thibodeau
16:12:07 [Zakim]
16:12:43 [sandro]
guest: Kai (Kai_) Eckert
16:13:25 [Luc]
@pgroth, is satya joining?
16:13:34 [pgroth]
I don't know
16:15:14 [pgroth]
@Luc I have to leave in 40 minutes
16:15:36 [Luc]
@pgroth OK
16:16:07 [Paolo]
what's happening?
16:16:28 [satya]
satya has joined #rdf-wg
16:16:30 [sandro]
zakim, who is on the call?
16:16:31 [Zakim]
On the phone I see Sandro, Paul_Groth, Scott_Bauer, kai_, AndyS, gavinc (muted), AZ, Luc, Paolo, davidwood
16:16:44 [Zakim]
16:17:02 [Zakim]
16:17:27 [MacTed]
Zakim, code?
16:17:27 [Zakim]
the conference code is 7334 (tel:+1.617.761.6200, MacTed
16:17:32 [davidwood]
16:17:33 [Zakim]
16:17:40 [MacTed]
Zakim, OpenLink_Software is temporarily me
16:17:40 [Zakim]
+MacTed; got it
16:17:42 [MacTed]
Zakim, mute me
16:17:43 [Zakim]
MacTed should now be muted
16:18:11 [Zakim]
16:18:37 [pgroth]
Scribe: Satya Sahoo
16:18:42 [sandro]
Guest: Satya Sahoo
16:19:07 [satya]
Luc: Introductions for the provenance WG
16:19:49 [satya]
Luc: After the first provenance WG F2F, started work on two working drafts
16:20:17 [satya]
Luc: 1. First working draft describes the provenance "conceptual model"
16:20:54 [satya]
Luc: 2. Second working draft describes encoding the provenance conceptual model in OWL2
16:21:26 [satya]
Luc: 3. Third working draft describes the protocol for accessing and querying provenance information
16:22:24 [Luc]
16:22:25 [satya]
Luc: In future, additional working drafts will be created including a Primer, XML encoding of the provenance conceptual model
16:22:53 [satya]
Luc: questions?
16:23:15 [satya]
David: The source of the requirement for encoding provenance model in XML?
16:23:33 [sandro]
luc: requirement for native XML serialization; some people want that; and JSON serialization. Non-RDF.
16:23:38 [satya]
Luc: There has been interest in non-RDF serialization
16:23:45 [gavinc]
I'm confused how needing a JSON seralization means we need an XML serialization
16:24:07 [sandro]
16:24:31 [sandro]
paul: plan to refer to group of triples by a URI
16:24:31 [satya]
David: can you please repeat your initial query
16:24:37 [pgroth]
16:24:40 [Luc]
@gavinc, this is a separate requirement, sorry for not being clear
16:24:50 [sandro]
16:25:01 [AndyS]
I heard that there was a need for straight XML (charter) and now Luc finds he needs JSON (as well, presumably)
16:25:13 [satya]
David: Is there a assumption that a group of RDF triples need to be referred to using an URI?
16:25:20 [Luc]
16:25:25 [satya]
Paul: Yes to David's question
16:25:31 [gavinc]
Ah, okay thanks AndyS
16:25:49 [Luc]
thanks, AndyS
16:25:55 [pgroth]
ack pgroth
16:25:56 [satya]
Paul: In query and access document, discusses use of URLs
16:25:58 [pgroth]
ack Luc
16:26:24 [davidwood]
I understand that provenance concerns resources other than RDF :)
16:27:01 [satya]
Luc: Uses an example scenario regarding an HTML document, the user tries to retrieve the provenance of the HMTL document
16:27:24 [satya]
Luc: Assume the use of HTTP for retrieving the provenance of the HTML document
16:28:00 [Luc]
16:29:02 [satya]
David: assumes that the resource and the provenance are distinct and can be independently accessed?
16:29:02 [sandro]
davidwood: Can I have provancence of an XL spreadsheet without modifying the spreadsheet?
16:29:42 [sandro]
gavin, I think you mean "can't *require* modifying...."
16:29:49 [gavinc]
errr, yes.
16:29:52 [satya]
Luc: There are several mechanism to access the provenance, a document may contain a reference to source to retrieve the provenance
16:29:55 [Luc]
16:29:56 [sandro]
gavin, Luc is sayin that's one option.
16:30:13 [davidwood]
RDF WG Charter:
16:30:34 [stain]
stain has joined #rdf-wg
16:30:45 [satya]
David: Describing RDF named graphs, as described in the RDF WG charter
16:31:11 [satya]
David: The description in the RDF WG charter is not definitive, will be decided by the WG
16:31:20 [davidwood]
RDF WG requirements: Define terminology in relation to named graphs
16:31:26 [davidwood]
"Standardize a model and semantics for multiple graphs and graphs stores"
16:32:04 [sandro]
davidwood: THere will be times we want to discuss the proveance of a singel triples, a group of triples, and times we dont' care.
16:32:11 [satya]
David: Provenance of a single triple and provenance of a group of triples
16:32:15 [satya]
@Sandro :)
16:32:55 [satya]
David: The teminology will be decided in future
16:32:59 [sandro]
16:33:03 [Luc]
16:33:10 [satya]
@Sandro: thanks! It helps, I keep falling behind
16:34:07 [satya]
Sandro: The proposal for named graph is minimal, it associated a URI to a group of triples (graph)
16:34:40 [satya]
Sandro: It was claimed that this met the requirements stated in the RDF WG charter
16:35:22 [davidwood]
Personally (chair hat off), I agree with Sandro on this.
16:35:33 [satya]
Sandro: Hope that concrete use cases can be shared by prov WG with RDF WG
16:37:04 [satya]
Luc: In SW community, there is a need to be able to make assertions about a group of triples and the RDF WG expects suggestions from prov WG
16:37:13 [sandro]
luc: "Provenance and Access Query" -- the problem of how folks give metadata to RDF data.
16:37:31 [sandro]
luc: ALSO, we need a way to scope provenance assertions that we've expressed as RDF.
16:37:48 [satya]
Luc: Need mechanisms to scope provenance assertions, named graphs may be a mechanism to support it
16:38:17 [davidwood]
q+ to ask how a provenance description can refer to a resource that doesn't have a URI.
16:38:25 [satya]
Sandro: Luc's example requirement is an advanced requirement for named graph
16:38:46 [gavinc]
Can someone write down that "First use case"?
16:39:22 [gavinc]
was it "there is a need to be able to make assertions about a group of triples" ?
16:39:55 [satya]
Luc: The first use case at, describes a scenario for retrieving provenance in RDF/XML format of a HTML document
16:40:25 [satya]
16:41:04 [pgroth]
ack sandro
16:41:05 [Luc]
ack sandro
16:41:11 [satya]
Sandro: Need to take a first pass on named graph problem before trying to tackle issue of scoping (?)
16:41:13 [sandro]
sandro: I suggest tacking the metadata problem (Luc 1) before the representation/serialization problem (Luc 2).
16:41:57 [satya]
David: Worried about some of Luc's assertions related to the provenance use case
16:42:29 [satya]
David: What happens if a resource may not have a URI associated with it
16:43:04 [MacTed]
q+ to say that the perfect seems to be blocking a starting point
16:43:15 [satya]
David: There is a divide in RDF WG - (a) all RDF comes from RDF databases (b) RDF comes from files on the Web
16:43:21 [davidwood]
ack davidwood
16:43:22 [Zakim]
davidwood, you wanted to ask how a provenance description can refer to a resource that doesn't have a URI.
16:43:35 [satya]
Paul: We would like to address both the issues
16:44:17 [satya]
Paul: An entity (file) is identified by a set of characteristics
16:44:57 [Luc]
16:45:06 [Luc]
ack satya
16:45:56 [sandro]
satya: "scoping" brings into question contextualizing, which I don't think the RDF/SemWeb folks have dealt with. There is an assumption that RDF is not related to context.
16:46:02 [sandro]
+1 agreed
16:46:10 [MacTed]
Zakim, unmte me
16:46:10 [Zakim]
I don't understand 'unmte me', MacTed
16:46:16 [MacTed]
Zakim, unmute me
16:46:16 [Zakim]
MacTed should no longer be muted
16:46:22 [Luc]
ack mact
16:46:22 [Zakim]
MacTed, you wanted to say that the perfect seems to be blocking a starting point
16:47:05 [sandro]
MacTed: I'm concerned that people are making very tight readings, when not warranted. downloading from web usually just mean viewing-in-browser.
16:47:21 [satya]
MacTed: Terminology used in description of use cases is not very specific, but is being interpreted in that sense
16:47:22 [davidwood]
+1 to MacTed. That's another reason we are being careful to define our terminology.
16:47:26 [sandro]
MacTed: There is a need to talk about a collected set of triples. "Named Graph". A subset of triples.
16:47:34 [sandro]
16:47:36 [sandro]
16:47:59 [sandro]
q+ to mention Graphs issues
16:48:06 [satya]
MacTed: There is a requirement to be able to refer to a collection of triples
16:48:13 [Luc]
16:48:47 [satya]
MacTed: Should be able to refer to set of triples consistently by different persons/users
16:49:19 [satya]
MacTed: There is communication gap between the RDF WG and prov WG and the call is help reconcile that
16:49:42 [sandro]
davidwood: The hope is RDF WG will publish it's chosen terms shortly.
16:49:47 [sandro]
... for public review
16:49:51 [satya]
David: Hope to define the terms in RDF WG
16:50:18 [satya]
MacTed: The lack of terminology in prov WG is blocking progress
16:50:27 [satya]
David: Similar issue in RDF WG
16:50:33 [Luc]
16:50:39 [Luc]
ack sandro
16:50:39 [Zakim]
sandro, you wanted to mention Graphs issues
16:50:50 [satya]
Sandro: Temporarily agreed on some terms to help progress
16:50:52 [pgroth]
16:51:10 [gavinc]
16:51:32 [gavinc]
The temporary terms
16:51:47 [sandro]
16:51:48 [satya]
Sandro: Gsnap and GBox terms defined in RDF WG
16:52:38 [satya]
Sandro: We need to be explicit about the blank nodes - whether they can be shared
16:52:42 [satya]
16:52:47 [davidwood]
I expect *any* resource to be ephemeral within some degree of time, so nothing we will do in the RDF WG or the Provenance WG will remove the 404 problem from the Web, nor do we need to try.
16:52:52 [Luc]
ack pgro
16:52:57 [sandro]
(issue-21 is shared-bnodes)
16:53:20 [gavinc]
16:53:20 [trackbot]
ISSUE-21 -- Can Node-IDs be shared between parts of a quad/multigraph format? -- open
16:53:20 [trackbot]
16:53:38 [satya]
Paul: can the provenance WG conceptual model document help the RDF WG arrive at common terminology
16:54:02 [davidwood]
q+ to ask whether the Provenance WG understands our g-* temporary terminology.
16:54:09 [satya]
Luc: Yes, we can explore this
16:54:32 [pgroth]
gotta go
16:54:36 [pgroth]
good luck everyone
16:54:39 [davidwood]
Thanks, Paul
16:54:56 [Zakim]
16:55:17 [satya]
Luc: Struggled in the provenance WG to define a term for an resource - e.g.: ability to refer to a file at a location with some content and the same file with different content
16:55:38 [Luc]
16:55:39 [satya]
Luc: Sandro may be referring to the same problem
16:55:42 [AndyS]
+1 to show and tell
16:56:01 [Luc]
16:56:06 [Luc]
ack saty
16:57:15 [satya]
Satya: What is the plan for sharing of blank nodes and association of semantics with it?
16:57:32 [satya]
Sandro: May modify the skolem function to address this
16:58:12 [satya]
David: Has the provenance WG understood the Gsnap, Gbox etc. being used by RDF WG
16:58:21 [davidwood]
ack davidwood
16:58:21 [Zakim]
davidwood, you wanted to ask whether the Provenance WG understands our g-* temporary terminology.
16:58:30 [satya]
Luc: No, we have not explored the terms
16:58:48 [satya]
@Luc: I have to leave
16:58:52 [satya]
16:59:13 [Zakim]
16:59:17 [sandro]
scribe: sandro
16:59:18 [Paolo]
leaving too, thank you
16:59:27 [Zakim]
16:59:34 [sandro]
luc: What are we doing next, procedurally?
16:59:35 [gavinc]
Luc: Wrapping up in the next 10 minutes, rather then talking about an example
16:59:45 [sandro]
davidwood: eg schedule next coordination call?
16:59:56 [sandro]
davidwood: After F2F, and after WDs.
17:00:06 [sandro]
davidwood: 12th & 13th of october.
17:00:08 [sandro]
17:00:16 [gavinc]
Could we get shared concrete examples BEFORE the next call?
17:00:22 [AndyS]
17:00:39 [davidwood]
17:00:45 [gavinc]
sandro: What input do we have?
17:01:16 [gavinc]
luc: two kinds of requirements, none of them are expressed at the level of detail mentioned, using temp graphs
17:01:27 [gavinc]
sandro: need use cases
17:02:02 [gavinc]
sandro: Thought we could get some use cases durring this call
17:02:31 [gavinc]
davidwood: Are there one or two use cases that you know we have?
17:02:39 [sandro]
sandro: I wanted uses cases before the F2F. Use cases like "communicating that Alice asserted certain triples (mutably, or immutably)".
17:02:50 [sandro]
zakim, who is on the call?
17:02:50 [Zakim]
On the phone I see Sandro, Scott_Bauer, kai_, AndyS, gavinc (muted), AZ, Luc, davidwood, MacTed
17:03:53 [Luc]
17:03:59 [sandro]
ted: Most of the use cases feel self evident, like what Sandro just said. An entity asserts triples; that collection needs to be referred to. If they have to be classed differently based on mutability, then they fall into one of two classes.
17:04:03 [sandro]
17:04:15 [sandro]
ack AndyS
17:04:19 [sandro]
17:05:42 [sandro]
AndyS: I think we need to get a lot more concrete on use cases. Some things hard to cover with g-star technology. Because we have two groups, and it's remote, we need VERY concrete cases, with real detail. Within that, decide on one or two we care about most, to be sure they can be done. Some risk that full set of requirements wont be met.
17:06:04 [sandro]
... I can see us taking too long if make it too broad, so let's focus on VERY concrete things that matter.
17:06:11 [davidwood]
17:06:17 [sandro]
luc: What's hard to map to g-star?
17:06:23 [sandro]
AndyS: yes, but no time right now.
17:06:55 [sandro]
AndyS: Show and tell could be good. Hard to tell the other side is bring precise.
17:07:30 [sandro]
public-rdf-prov list?
17:07:43 [AndyS]
ack me
17:08:12 [sandro]
luc: Ted and Sandro write it down and let Prov-WG review it.
17:09:10 [kai_]
q+ to ask for bridge persons between the two groups
17:09:58 [sandro]
kai_: We're looking for bridge group.
17:10:19 [Luc]
ack kai
17:10:19 [Zakim]
kai_, you wanted to ask for bridge persons between the two groups
17:11:00 [sandro]
davidwood: Can we agree to only refer to things by URI ?
17:11:14 [davidwood]
ack me
17:12:03 [sandro]
luc: Ultimately, we want to talk about provenance of resources, but we don't jhave the final story. Using URIs is definitely one of our hypotheses.
17:12:12 [kai_]
I ask people who are interested to help in the communication between the two groups to contact me.
17:12:18 [Luc]
17:12:20 [sandro]
luc: davidwood can you ask this to the new mailing list?
17:12:31 [sandro]
kai_, why not just ask themt o join the list?
17:12:36 [MacTed]
from my perspective, Prov is *not* only interested in provenance of "web" resources nor "resources" per se -- but *entities*
17:13:04 [kai_]
@sandro works, too :-)
17:13:05 [davidwood]
MacTed, what is an entity?
17:13:13 [MacTed]
a thing which can be named
17:13:26 [gavinc]
Named with what? ;)
17:13:32 [davidwood]
Named via a URI? ;)
17:13:33 [gavinc]
say ... an IRI? ;)
17:13:45 [davidwood]
17:14:28 [Luc]
17:14:38 [AZ]
17:14:42 [Zakim]
17:14:43 [MacTed]
preferably, yes -- IRI. my concern is not to limit to "resource" nor "web resource" -- because when docs are written that way, people take them to be not viable for concrete things
17:14:43 [Zakim]
17:14:44 [Zakim]
17:14:45 [Zakim]
17:14:46 [Zakim]
17:14:46 [Zakim]
17:14:47 [Zakim]
17:14:50 [Zakim]
17:14:54 [Zakim]
17:14:55 [Zakim]
SW_RDFWG(GraphsTF)12:00PM has ended
17:14:56 [MacTed]
(where "concrete" means "not network transmissible")
17:14:57 [Zakim]
Attendees were Sandro, Scott_Bauer, Paul_Groth, kai_, AndyS, gavinc, AZ, Luc, Paolo, davidwood, Yolanda, Satya_Sahoo, MacTed
17:15:41 [davidwood]
MacTed, REST allows *anything* to be a resource.
17:15:51 [davidwood]
…and thus named
17:16:12 [sandro]
ie Ted is talking about "Non-Information-Resources"
17:17:10 [davidwood]
I know - "anything" was meant explicitly to include non-information resources
17:17:20 [AndyS]
REST does not name what is accessed (representation).
17:17:28 [MacTed]
davidwood - yes, it does. but when examples are written only around "Information Resources", they are often interpreted as so specific...
17:21:04 [davidwood]
MacTed, sure. I guess we'll have to write good examples then.
17:22:13 [davidwood]
AndyS, yes. That's why I still hold out hope that we can treat URIs as names. If you resolve a UR*L* to get a representation, you change its URL if you save that representation to a file system or database.
17:22:59 [AndyS]
c.f. 3Store (except the name is a bNode)
17:23:59 [davidwood]
Skolemized? :)
17:24:33 [AndyS]
I think a good thing we can do is to get this naming (UUID URIs) sorted out. It's not the only approach though - graphs literals is also workable theory but may be less of a way forward
17:24:39 [AndyS]
17:25:19 [AndyS]
Def can't have too much skolemization. It's like indirection but without the direction.
17:26:35 [davidwood]
17:26:53 [davidwood]
Yes, I agree that getting naming sorted would allow us a way forward.
17:27:42 [davidwood]
Sorry - I need to run.
17:27:47 [AndyS]
Well, that's F2F filled. Now, world peace and climate change.
17:32:35 [LeeF]
LeeF has joined #rdf-wg
17:33:59 [trackbot]
trackbot has joined #rdf-wg
18:14:27 [mischat_]
mischat_ has joined #rdf-wg
19:12:03 [mischat]
mischat has joined #rdf-wg
19:59:34 [Zakim]
Zakim has left #rdf-wg