IRC log of rdf-wg on 2013-06-19

Timestamps are in UTC.

15:00:52 [RRSAgent]
RRSAgent has joined #rdf-wg
15:00:52 [RRSAgent]
logging to
15:00:54 [trackbot]
RRSAgent, make logs world
15:00:54 [Zakim]
Zakim has joined #rdf-wg
15:00:56 [trackbot]
Zakim, this will be 73394
15:00:56 [Zakim]
ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start now
15:00:57 [trackbot]
Meeting: RDF Working Group Teleconference
15:00:57 [trackbot]
Date: 19 June 2013
15:01:04 [ivan]
zakim, dial ivan-voip
15:01:04 [Zakim]
ok, ivan; the call is being made
15:01:24 [ivan]
Chair: David Wood
15:01:29 [ivan]
zakim, who is here?
15:01:29 [Zakim]
SW_RDFWG()11:00AM has not yet started, ivan
15:01:30 [Zakim]
On IRC I see RRSAgent, pfps, gkellogg, tbaker, PatH, pchampin, manu, Guus, AndyS, TallTed, SteveH, davidwood, ivan, manu1, gavinc, yvesr, sandro, trackbot, ericP, mischat
15:01:30 [davidwood]
Zakim? I am here...
15:02:05 [TallTed]
TallTed has changed the topic to: RDF WG -- -- current agenda
15:02:07 [AZ]
AZ has joined #rdf-wg
15:02:19 [TallTed]
Zakim, this is 73394
15:02:19 [Zakim]
ok, TallTed; that matches SW_RDFWG()11:00AM
15:02:22 [Zakim]
15:02:23 [TallTed]
Zakim, who's here?
15:02:24 [Zakim]
On the phone I see +1.408.992.aaaa, ??P24, davidwood, Ivan, PatH, [IPcaller]
15:02:24 [Zakim]
On IRC I see AZ, Zakim, RRSAgent, pfps, gkellogg, tbaker, PatH, pchampin, manu, Guus, AndyS, TallTed, SteveH, davidwood, ivan, manu1, gavinc, yvesr, sandro, trackbot, ericP,
15:02:27 [Zakim]
... mischat
15:02:28 [AndyS]
zakim, IPCaller is me
15:02:28 [Zakim]
+AndyS; got it
15:02:40 [gkellogg]
zakim, I am ??P24
15:02:43 [Zakim]
+gkellogg; got it
15:02:53 [Zakim]
15:03:26 [TallTed]
Zakim, OpenLink_Software is temporarily me
15:03:27 [Zakim]
+TallTed; got it
15:03:29 [Zakim]
15:03:35 [TallTed]
Zakim, mute me
15:03:35 [Zakim]
TallTed should now be muted
15:03:44 [Zakim]
+ +081165aabb
15:03:52 [AZ]
Zakim, aabb is me
15:03:52 [Zakim]
+AZ; got it
15:04:13 [Zakim]
15:06:34 [gkellogg]
15:06:54 [markus]
markus has joined #rdf-wg
15:07:36 [AndyS]
scribenick: AndyS
15:07:43 [AndyS]
scribe: AndyS
15:07:49 [davidwood]
PROPOSED to accept the minutes of the 12 June telecon:
15:07:52 [AndyS]
15:07:57 [pfps]
minutes are fine
15:08:24 [davidwood]
RESOLVED to accept the minutes of the 12 June telecon:
15:08:30 [ericP]
davidwood, i'm in the LDP F2F, but can switch over here when talking about Turtle
15:08:35 [davidwood]
Review of action items
15:08:36 [davidwood]
15:08:36 [davidwood]
15:08:41 [AndyS]
topic: Action items
15:09:13 [davidwood]
15:09:13 [trackbot]
ACTION-226 -- Richard Cyganiak to implement ISSUE-111 resolution -- due 2013-02-13 -- OPEN
15:09:13 [trackbot]
15:09:24 [AndyS]
pfps: Close action on ACTION-226 for ISSUE-111
15:09:36 [AndyS]
... no-op.
15:10:12 [AndyS]
Close ACTION-226
15:10:12 [trackbot]
Closed ACTION-226 Implement ISSUE-111 resolution.
15:10:53 [davidwood]
15:10:54 [trackbot]
ACTION-256 -- Gavin Carothers to link TriG to new text in RDF Concepts -- due 2013-05-01 -- OPEN
15:10:54 [trackbot]
15:10:55 [AndyS]
pfps: Need work on ACTION-256
15:11:11 [ericP]
ACTION: ericP to respond to LC issues (11, 12, 18, 37)
15:11:12 [trackbot]
Created ACTION-273 - Respond to LC issues (11, 12, 18, 37) [on Eric Prud'hommeaux - due 2013-06-26].
15:11:39 [Zakim]
15:11:40 [ericP]
that should close 271
15:11:42 [markus]
zakim, ??P35 is me
15:11:42 [Zakim]
+markus; got it
15:11:58 [AndyS]
(Archaeology occurs)
15:13:26 [Zakim]
15:13:27 [gavinc]
15:13:42 [sandro]
sandro: I see the resolution to issue-131 has not yet been reflected in rdf-concepts.
15:13:47 [Souri]
Souri has joined #rdf-wg
15:13:49 [Zakim]
15:14:00 [AndyS]
pfps: action-107 has not made it into concepts.
15:14:05 [AndyS]
15:14:05 [trackbot]
ACTION-107 -- Richard Cyganiak to add a note to RDF Concepts re ISSUE-75 -- due 2011-10-20 -- CLOSED
15:14:05 [trackbot]
15:14:09 [Guus]
zakim, +[IPcaller] is me
15:14:09 [Zakim]
sorry, Guus, I do not recognize a party named '+[IPcaller]'
15:14:30 [Guus]
zaki, IPCaller is me
15:14:35 [AndyS]
gavinc: wrong action?
15:15:00 [Guus]
zakim, IPCaller is me
15:15:00 [Zakim]
+Guus; got it
15:15:12 [Guus]
zakim, mute me
15:15:12 [Zakim]
Guus should now be muted
15:15:27 [AndyS]
pfps: 24 April ... blank nodes can be shared
15:15:41 [AndyS]
gavinc: done ... need more for blank node for graph names.
15:15:43 [davidwood]
15:16:15 [AndyS]
davidwood: I'll do that
15:16:41 [davidwood]
ACTION: davidwood to implement in Concepts
15:16:41 [trackbot]
Created ACTION-274 - Implement in Concepts [on David Wood - due 2013-06-26].
15:17:18 [AndyS]
pfps: success on action-272
15:18:11 [PatH]
noisy typing
15:18:16 [AndyS]
davidwood: proposal to deal with public comments which are discussion and formal comments
15:18:16 [pfps]
action-272 is closed
15:18:29 [sandro]
close action-272
15:18:30 [trackbot]
Closed ACTION-272 Send email describing the differences between Antoine's view and the current draft (eg you don't know what the datatype interpretation is).
15:18:31 [Guus]
for the record, I closed action 271, done
15:19:02 [AndyS]
... list is part of the formal process but it has had "other stuff" recently (and before). Need to get under control.
15:19:26 [AndyS]
... danger of loosing (real) comments
15:19:35 [pfps]
15:19:41 [TallTed]
15:19:50 [AndyS]
... so please do not reply until there is a formal thing to say as decided by a chair.
15:20:10 [pfps]
q- because Pat said what I wanted to
15:20:15 [davidwood]
ack pfps
15:20:25 [AndyS]
path: std response for incoming.
15:20:34 [AndyS]
... will draft text
15:20:42 [AndyS]
sandro: set up as autoreply
15:21:13 [AndyS]
gavinc: where should be the discussion be?
15:21:28 [Guus]
semantic-web list is the propoer forum for many discussions
15:21:46 [AndyS]
davidwood: else where eg. ... need to have a functioning comments process
15:21:52 [Guus]
publi-comments only for relevant design dcomments
15:21:53 [pfps]
+1 to this sentiment
15:22:45 [PatH]
Proposed autorespond text: DUring the Last Call period, this email list is restricted to change requests on the LC documents and official responses to those requests. Members of the WG will not respond to more general comments or discussions on this list. PLease re-post your comment on a different public list. THank you.
15:22:54 [AndyS]
gavinc: e.g. discussion on turtle as has happened.
15:23:04 [pfps]
15:23:29 [pfps]
the autorespond text should be permanent, not just during last call
15:23:47 [AndyS]
Extra list? rdf-spec-discuss? Specifically NOT comments.
15:23:51 [davidwood]
ack pfps
15:24:19 [Guus]
in other groups we have always done it this way: first propose a draft response to the WG
15:24:27 [PatH]
SO lets add a pointer to that list in the autrespond text.
15:24:29 [AndyS]
pfps: semantic-web@ is for this.
15:25:15 [AndyS]
... drive-by email lists
15:25:20 [davidwood]
15:25:59 [TallTed]
Zakim, unmute me
15:25:59 [Zakim]
TallTed should no longer be muted
15:26:00 [AndyS]
davidwood: pat/text :: sandro/setup auto :: davidwood/tell list
15:26:05 [Guus]
the message to the chairs should normally be cc or to WG as a whole, so all can see what;s happening
15:26:46 [AndyS]
sandro: TallTed -- please pass on this conversation to Kingsley.
15:26:53 [pchampin]
pchampin has joined #rdf-wg
15:27:06 [AndyS]
TallTed: OK - will try.
15:27:31 [TallTed]
Zakim, mute me
15:27:31 [Zakim]
TallTed should now be muted
15:27:35 [AndyS]
topic: LC Drafts of Concepts and Semantics
15:27:58 [AndyS]
davidwood: semantics first
15:28:06 [davidwood]
Discussion thread on review by Antoine of Semantics:
15:28:06 [davidwood]
15:28:34 [AndyS]
(scribe has not followed this thread in email in all details)
15:29:05 [AndyS]
zakim, who is on the phone?
15:29:05 [Zakim]
On the phone I see +1.408.992.aaaa, gkellogg, davidwood, Ivan, PatH, AndyS, TallTed (muted), Sandro, AZ, GavinC, markus, Souri, Guus (muted)
15:30:09 [Zakim]
15:30:12 [ivan]
15:30:16 [pchampin]
zakim, ??P5 is me
15:30:16 [Zakim]
+pchampin; got it
15:30:17 [AndyS]
AZ: datatype URIs -- new design has just a set of IRIs , no maps -- we need to know what they denote
15:30:44 [AndyS]
... second issue is entailment of set sof RDF graphs (scribe: union discussion?)
15:31:47 [davidwood]
15:31:50 [AndyS]
... set entails the union, it was each of the sets. Does not follow normal practice. RDF is not the same as conjunction (correct?)
15:32:26 [AndyS]
... will provide some proposed text that meets my concerns.
15:32:50 [davidwood]
ack ivan
15:33:10 [PatH]
15:34:23 [AndyS]
ivan: restrictions on datatype URIs ... normative text around it covers the conditions ... semantic conditions are not all in the maths
15:35:02 [AndyS]
AZ: pfps argues there is not real difference as RDF 2004. But then I don't don't understand why the new design/ expression of design is better.
15:35:07 [sandro]
It's editorial, but important editorially.
15:35:28 [pfps]
there may be a point on the definition of entailment on sets of graphs
15:35:34 [AndyS]
ivan: better exposition
15:35:59 [Guus]
+1 to Ivan
15:36:05 [AndyS]
.. a big issue has been the limted readership of the doc
15:36:46 [AndyS]
AZ: I am not saying it is the same design.
15:37:17 [Guus]
ack PatH
15:38:04 [AndyS]
path: latest draft has some improvements - let's discuss relative to that
15:38:14 [AndyS]
... fixes a small point
15:38:31 [AZ]
15:38:53 [AndyS]
... a slight difference to 2004 is that 2004 D-mapping allows xsd:string mapping to data times
15:39:18 [AndyS]
... and now it is not allowed to have a different D-entailment
15:39:30 [Guus]
Note that LC for Semantics is on our critical path; we have to have very good reasons NOT to go there; Feature at Risk is a possibility I guess
15:39:33 [AndyS]
... also I belive/hope that doc is more accessible.
15:39:36 [pfps]
not only is it illegal to mess with, e.g, xs datatypes, the mechanism in RDF is now consonant with the mechanism in xs datatypes
15:40:05 [AndyS]
AZ: can add a constraint that these IRIs must map to the normal meaning
15:40:28 [AndyS]
path: but then no need for a map
15:40:38 [AndyS]
AZ: custom datatypes
15:40:57 [pfps]
15:41:06 [Guus]
ack AZ
15:41:30 [AndyS]
path: can only recognize IRIs that are datatypes
15:42:28 [AndyS]
... IF e.g. GET shows that it is a datatype, then must use that defn of the datatype.
15:42:30 [gavinc]
"RDF processors which are not able to determine which datatype is identifier by an IRI cannot recognize that IRI, and should treat any literals type with that IRI as unknown names." this is bit we are talking about?
15:43:53 [davidwood]
15:44:02 [Guus]
Guus has joined #rdf-wg
15:44:29 [davidwood]
ack pfps
15:44:59 [PatH]
gavinC, yes.
15:45:08 [AndyS]
(discussion of the true purpose of the MT doc)
15:45:41 [gavinc]
btw, I like the language on what to do when there is an unknown datatype in the new semantics document then the old one. I had no clue what to do with unknown datatypes based on the old one, and the new one is VERY clear and even uses MAY/SHOULD. It's MUCH better
15:46:16 [AZ]
The normal Web machinery is not existing in the current version either
15:48:10 [davidwood]
Gentlemen, I suggest you use the queue to respond. It will assist in limiting emotion.
15:49:20 [Guus]
+1 to peter
15:50:55 [AndyS]
davidwood: Thank you for the discussion. AZ - do you need time to work on a proposal.
15:51:17 [ivan]
15:51:31 [AndyS]
AZ: easier for me to work on a proposal for datatypes. Will try to do it quite quickly.
15:51:36 [pfps]
15:51:43 [AndyS]
... for other issue - less clear.
15:52:03 [AndyS]
davidwood: can we go to LC?
15:52:05 [davidwood]
ack ivan
15:52:14 [AndyS]
(someone) no
15:52:45 [PatH]
+1 Ivan
15:53:47 [pfps]
Antoine appears to feel that there is something wrong in the current situation, so I don't think that agrees that it is just stylistic
15:54:18 [AndyS]
AZ: This is not me raising a new issue.
15:54:42 [Guus]
antoine is right about the timing, but we're just not overcoming this disagreement
15:56:35 [pfps]
the question here, I think, is whether there is something technically wrong in the current semantics
15:57:08 [pfps]
q+ to talk about entailment
15:57:39 [Guus]
zakim, unmute me
15:57:39 [Zakim]
Guus should no longer be muted
15:57:53 [davidwood]
ack pfps
15:57:53 [Zakim]
pfps, you wanted to talk about entailment
15:59:03 [AZ]
Not defined in term of merge
15:59:17 [AndyS]
pfps; I now see the point about in the "union" strand of discussion
15:59:18 [Guus]
zakim, mute me
15:59:18 [Zakim]
Guus should now be muted
15:59:29 [AZ]
defined according to traditional conjunction
16:00:11 [manu]
zaim, code?
16:00:16 [manu]
zakim, code?
16:00:16 [Zakim]
the conference code is 73394 (tel:+1.617.761.6200, manu
16:00:26 [Zakim]
16:00:33 [PatH]
16:00:34 [manu]
zakim, I am ??P53
16:00:35 [Zakim]
+manu; got it
16:01:44 [AndyS]
pfps: example - split one graph in two , share bnode across the halfs.
16:02:03 [AZ]
16:02:06 [AZ]
16:02:16 [AndyS]
... union rejoins the use of blank nodes ... merge looses the connection in the original big graph.
16:03:38 [pchampin]
16:03:45 [pchampin]
16:04:38 [davidwood]
ack PatH
16:04:43 [gavinc]
can't we just lean on datasets? Where we SAY that blank nodes are shared inside a dataset?
16:04:46 [pfps]
entailments of sets of RDF graphs - 2004 effectively uses merge 2013 uses union
16:05:17 [pfps]
16:05:32 [AndyS]
path: because we have said we will support shared bnodes then merge would make this meaningless.
16:06:48 [AndyS]
... lots of pictures in the current doc to explain this all.
16:06:51 [sandro]
Ahhhh. Yes, Merge is kind of .... wacky ... now. There should just be UNION and also a COPY_GRAPH operation the replace blank nodes with a fresh ones.
16:06:55 [davidwood]
ack AZ
16:07:02 [pchampin]
16:07:09 [pfps]
I don't agree with this analysis by Pat - people will not determine our stance on combining graphs by the definition of entailment
16:07:16 [davidwood]
16:07:19 [PatH]
q+ again
16:08:06 [AndyS]
AZ: if a system knows two bnodes are the same they can use union
16:09:03 [Souri]
Then, is the following correct? (users decide which of the following to use when combining) UNION => reuse bNode labels, MERGE => generate new unique bNode labels
16:09:44 [pchampin]
+1 Pat
16:09:49 [ivan]
+1 to Pat
16:09:50 [AndyS]
Example -- :mary :hasChild _:b . :john :hasChild _:b .
16:10:13 [AndyS]
path: in that graph - mary and john have a child in common.
16:10:33 [pchampin]
AZ, thought experiment: does a graph entail the same thing as the union of all its triples (considered as singleton graphs)?
16:10:50 [AndyS]
AZ: in subgraphs {:mary :hasChild _:b } { :john :hasChild _:b }
16:11:02 [AZ]
it's not a bug!
16:11:12 [AndyS]
Path: bug is that the idea of bnode scope is not in the definitions.
16:11:34 [Zakim]
16:11:49 [Zakim]
16:11:57 [ericP]
Zakim, [GVoice] is me
16:11:57 [Zakim]
+ericP; got it
16:12:24 [pfps]
this actually has little to do with blank nodes per se, as the same issue arises with free variables in FOL formulae if you interpret free variables existentially
16:13:00 [pchampin]
the question is: where is the quantifier of those existential variables?
16:13:02 [davidwood]
Zakim, mute ericP
16:13:02 [Zakim]
ericP should now be muted
16:13:06 [tlr]
tlr has joined #rdf-wg
16:13:25 [pfps]
q+ to say that blank nodes *do* work like variables in quantifier-free first-order formulae
16:13:27 [davidwood]
16:13:28 [pchampin]
AZ seems to assume that it is at the graph level
16:14:01 [pchampin]
by deciding that graph can share blank nodes, my interpretation is that we decided that it could be at a higher level
16:14:21 [pchampin]
16:14:33 [tlr]
guys, can you get off the bridge?
16:14:35 [AndyS]
You can't share bNodes unless they originate from the same place because of all syntax parsing havign scoped labels.
16:14:41 [tlr]
we're having other groups that can't get their meetings running
16:14:44 [AndyS]
davidwood: time
16:14:48 [Zakim]
16:14:49 [davidwood]
ack pfps
16:14:49 [Zakim]
pfps, you wanted to say that blank nodes *do* work like variables in quantifier-free first-order formulae
16:14:49 [pfps]
take a formula P(x) and another formula Q(x), when you form their conjunction you don't change the variables, you just put an "and" over them (i.e., union them)
16:15:16 [davidwood]
ack again
16:16:00 [AndyS]
Next time can we progress the TriG Turtle NQ, NT which have slipped again.
16:16:12 [tlr]
please kill this meeting now
16:16:16 [PatH]
I guess I feel that I have explained my posiiton as clearly and as fully as I can, and that the current draft also explains it as clearly as I know how. I rest my case.
16:16:19 [tlr]
there are plenty of other calls having trouble right now
16:16:23 [AndyS]
Should be "easy" to agree to go to publication or find a process to get there.
16:16:26 [Zakim]
16:16:49 [Zakim]
16:16:52 [AndyS]
davidwood: I have worked through pfps review of concepts
16:16:52 [pchampin]
16:16:55 [Zakim]
16:16:56 [Zakim]
16:16:56 [Zakim]
- +1.408.992.aaaa
16:16:56 [Zakim]
16:16:57 [Zakim]
16:16:57 [Zakim]
16:16:57 [AndyS]
16:16:58 [Zakim]
16:17:00 [Zakim]
16:17:01 [Zakim]
16:17:03 [Zakim]
16:17:20 [davidwood]
tlr, understood and adjourned
16:17:32 [Zakim]
16:17:37 [Zakim]
16:17:43 [Zakim]
SW_RDFWG()11:00AM has ended
16:17:44 [Zakim]
Attendees were +1.408.992.aaaa, Ivan, davidwood, PatH, AndyS, gkellogg, TallTed, Sandro, +081165aabb, AZ, GavinC, markus, Souri, Guus, pchampin, manu, ericP
16:20:14 [tlr]
16:22:09 [AndyS]
sandro? W3C team?
16:22:20 [sandro]
16:22:21 [AndyS] --> internal server error
16:22:26 [sandro]
hold on
16:22:32 [AndyS]
loosing my edits :-(
16:22:44 [sandro]
16:22:45 [davidwood]
16:23:45 [sandro]
the internal error was, "Can't parse chatlog line '16:13:06 tlr has joined #rdf-wg'" and it's my quick-and-dirty code that made that result in lossage.
16:24:16 [AndyS]
I had to edit tlr lines because he's not in the WG.
16:24:30 [AndyS]
If i avoid that, can I continue editting?
16:24:52 [sandro]
I believe so.
16:25:20 [sandro]
same error
16:25:32 [AndyS]
No - can't click on the date link to get a fresh copy to edit.
16:26:06 [AndyS]
seems like its written before checking for parsability
16:27:11 [sandro]
hold on.
16:27:55 [AndyS]
antici... pation.
16:28:01 [sandro]
RRSAgent, bye
16:28:01 [RRSAgent]
I see 2 open action items saved in :
16:28:01 [RRSAgent]
ACTION: ericP to respond to LC issues (11, 12, 18, 37) [1]
16:28:01 [RRSAgent]
recorded in
16:28:01 [RRSAgent]
ACTION: davidwood to implement in Concepts [2]
16:28:01 [RRSAgent]
recorded in