IRC log of sparql on 2011-06-28

Timestamps are in UTC.

13:54:24 [RRSAgent]
RRSAgent has joined #sparql
13:54:24 [RRSAgent]
logging to http://www.w3.org/2011/06/28-sparql-irc
13:54:26 [trackbot]
RRSAgent, make logs world
13:54:26 [Zakim]
Zakim has joined #sparql
13:54:28 [trackbot]
Zakim, this will be 77277
13:54:28 [Zakim]
ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 6 minutes
13:54:28 [LeeF]
zakim, this will be SPARQL
13:54:29 [trackbot]
Meeting: SPARQL Working Group Teleconference
13:54:29 [trackbot]
Date: 28 June 2011
13:54:29 [Zakim]
ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 6 minutes
13:56:52 [Zakim]
SW_(SPARQL)10:00AM has now started
13:56:55 [Zakim]
SW_(SPARQL)10:00AM has ended
13:56:56 [Zakim]
Attendees were
13:57:21 [kasei]
heh
13:57:48 [SteveH]
well, that was a quick call, see you all next week :)
13:58:21 [Zakim]
SW_(SPARQL)10:00AM has now started
13:58:30 [Zakim]
+kasei
13:59:27 [Zakim]
+??P6
13:59:33 [cbuilara]
zakim, ??P6 is me
13:59:33 [Zakim]
+cbuilara; got it
13:59:39 [Zakim]
+caro
13:59:57 [Zakim]
+LeeF
14:00:11 [Zakim]
+ +44.189.583.aaaa
14:00:13 [LeeF]
zakim, who's on the phone?
14:00:14 [Zakim]
On the phone I see kasei, cbuilara, caro, LeeF, +44.189.583.aaaa
14:00:38 [NickH]
Zakim, +44.189.583.aaaa is me
14:00:38 [Zakim]
+NickH; got it
14:00:44 [NickH]
me!
14:00:48 [Olivier]
Olivier is on the phone
14:00:50 [Zakim]
+??P14
14:00:56 [SteveH]
Zakim, ??P14 is me
14:00:57 [Zakim]
+SteveH; got it
14:00:59 [LeeF]
zakim, caro is Olivier
14:01:00 [Zakim]
+Olivier; got it
14:01:16 [LeeF]
Regrets: Axel, Alex, Paul
14:01:50 [LeeF]
zakim, who's on the phone?
14:01:51 [Zakim]
On the phone I see kasei, cbuilara, Olivier, LeeF, NickH, SteveH
14:02:05 [NickH]
sure
14:02:11 [LeeF]
scribenick: NickH
14:02:16 [LeeF]
Chair: LeeF
14:02:21 [Zakim]
+??P21
14:02:30 [AndyS]
zakim, ??P21 is me
14:02:30 [Zakim]
+AndyS; got it
14:03:27 [LeeF]
topic: Admin
14:03:33 [LeeF]
PROPOSED: http://www.w3.org/2009/sparql/meeting/2011-06-21
14:04:22 [NickH]
LeeF: most of our time last week was spent talking about the RDF-WG decision on the literal datatype
14:04:46 [LeeF]
RESOLVED: Accept last week's minutes from http://www.w3.org/2009/sparql/meeting/2011-06-21
14:04:49 [AndyS]
Mins Look OK to me
14:05:21 [NickH]
LeeF: we have a couple of acknowledgents to comment responses
14:05:52 [NickH]
LeeF: did carlo is the comment about a typo
14:05:58 [NickH]
carlo: yes
14:06:10 [kasei]
still waiting on approval to send out RV-5
14:06:24 [NickH]
LeeF: two new comments
14:07:35 [AndyS]
IH-2 is minor and will get done sometime. OWL-Semantics ==> OWL 2 Semantics
14:07:36 [NickH]
LeeF: first is comment NL-1
14:07:40 [LeeF]
http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Jun/0014.html
14:09:23 [NickH]
LeeF: when we were doing the shortcuts for SPARQL Update, you can leave out the template from the CONSTRUCT verb
14:09:48 [NickH]
LeeF: and it just copies the template from the WHERE clause
14:10:01 [AndyS]
Yes - that's the defn - might (haven't investigated) work as template = all triple patterns mentions for UNION + OPTIONAL
14:10:05 [NickH]
LeeF: concerns about OPTIONAL and UNION being allowed
14:11:00 [AndyS]
http://service.demo.lagrummet.se/view/publ/sfs/1991:1469?query
14:11:14 [NickH]
AndyS: it would be pretty hairy to support FILTER not bound
14:11:58 [SteveH]
I think it's just too hard
14:12:04 [NickH]
AndyS: it would have to say something about all the triples that are posively touched
14:12:21 [NickH]
LeeF: is there a concept that would be reusable in the current doucment?
14:12:47 [NickH]
LeeF: I am inclined to politely decline - out of time and out of scope
14:13:01 [NickH]
AndyS: too late to put in new features like this
14:14:04 [NickH]
LeeF: I will take ownership of the response
14:15:00 [NickH]
LeeF: the other comment was JD-6
14:15:24 [chimezie]
chimezie has joined #sparql
14:17:11 [Zakim]
+chimezie
14:17:12 [chimezie]
chimezie has joined #sparql
14:17:38 [chimezie]
hey (sorry I'm late)
14:17:44 [chimezie]
Zakim, who is here?
14:17:44 [Zakim]
On the phone I see kasei, cbuilara, Olivier, LeeF, NickH, SteveH, AndyS, chimezie
14:17:47 [Zakim]
On IRC I see chimezie, Zakim, RRSAgent, cbuilara, Olivier, LeeF, AndyS, SteveH, iv_an_ru_, pgearon, ericP, sandro, trackbot, kasei, NickH, alepas
14:17:48 [NickH]
AndyS: if we are now going to allow functions like SUM to return strings, it gets more complex
14:18:04 [AndyS]
SUM(my:fixTheData(?x)) where my:fixTheData(?x) gets things into being a number
14:18:11 [Zakim]
-SteveH
14:18:16 [NickH]
AndyS: restricting aggregate functions to numbers feels safer
14:18:36 [kasei]
q+
14:18:38 [Zakim]
+??P14
14:18:42 [SteveH]
Zakim, ??P14 is me
14:18:42 [Zakim]
+SteveH; got it
14:19:00 [SteveH]
q+
14:19:03 [LeeF]
ack kasei
14:19:09 [NickH]
LeeF: his goal is to let the query engine automatically sort it out, rather than the query writer
14:19:31 [LeeF]
ack SteveH
14:19:54 [MattPerry]
MattPerry has joined #sparql
14:20:24 [NickH]
SteveH: this is a straight forward trade-off between easy of use and safety. When it comes to integers, I would edge on the side of safety.
14:20:45 [AndyS]
q+ to note this does not *require* SUM to be def'ed as +
14:20:58 [Zakim]
+MattPerry
14:21:04 [kasei]
q+
14:21:31 [LeeF]
http://www.w3.org/TR/sparql11-query/#operatorExtensibility
14:22:14 [kasei]
q-
14:22:53 [NickH]
SteveH: we don't have a paragraph of text that says it is valid to extend aggregate functions
14:23:20 [AndyS]
We don't say (and maybe should) that functions (and aggs) can be extended where error -- we did in SPARQL 1.0 (editorial change)
14:23:39 [LeeF]
ack AndyS
14:23:39 [Zakim]
AndyS, you wanted to note this does not *require* SUM to be def'ed as +
14:24:04 [NickH]
AndyS: in Sparql 1.0 we say that you can extend any function in the operator table
14:24:32 [NickH]
AndyS: at the moment we don't explictly say that you can extend functions that would other cause an error
14:25:06 [SteveH]
+1 to AndyS
14:25:11 [NickH]
AndyS: if we changed SUM() to mean the same thing as +, that would be a big change
14:25:16 [LeeF]
q?
14:25:34 [NickH]
AndyS: he just has rubbish data that he wants to clean up!
14:26:07 [NickH]
SteveH is the owner and has written a draft response
14:26:20 [AndyS]
And "1+1"^^xsd:string ?
14:26:29 [NickH]
LeeF: any more comments on comments?
14:26:47 [LeeF]
topic: Graph terminology telecon
14:26:57 [NickH]
LeeF: next topic is joint graph terminology comment from last week
14:27:07 [SteveH]
http://www.w3.org/2009/sparql/wiki/CommentResponse:JB-6 <- now with extra para
14:27:46 [NickH]
chimezie: G-Snap is roughly an RDF Graph
14:28:00 [NickH]
chimezie: G-Box is is RDF Graph Content
14:28:11 [NickH]
chimezie: G-Box is an RDG Graph Document
14:28:20 [NickH]
chimezie: G-Box is an RDF Graph Document
14:28:52 [NickH]
chimezie: we didn't get to the point of actually discussing the text
14:29:17 [NickH]
LeeF: there was some some conern about some of the words used
14:29:22 [NickH]
chimezie: there are two concerns
14:29:38 [NickH]
chimezie: the notion of an RDF Graph is not formally defined
14:29:56 [NickH]
chimezie: they would prefer a different term than RDF Document
14:31:16 [NickH]
LeeF: this is pretty easy as far as the SPARQL WG is concerned
14:31:59 [NickH]
chimezie: if those actions get to the point where they want to change text in mature documents
14:32:14 [NickH]
LeeF: probably had to play it by ear
14:32:55 [NickH]
LeeF: do you think that the actions will result in significant changes to doucment?
14:33:20 [NickH]
chimezie: I don't think it will result in a wholesale rewrite
14:33:36 [NickH]
AndyS: are there going to be minutes and a summary?
14:33:59 [NickH]
LeeF: Richard got hold of IRC logs and is going to create minutes and a summary
14:34:19 [NickH]
LeeF: I have a black market copy of the IRC logs, which is available for the right price
14:34:51 [NickH]
LeeF: I will send it to the list but Richard will be creating a nicely formatted version
14:35:05 [chimezie]
Zakim, mute me
14:35:05 [Zakim]
chimezie should now be muted
14:35:07 [LeeF]
topic: RDF WG decision on simple literals
14:35:19 [NickH]
LeeF: wish sandro was here for the last topic on the agenda
14:36:18 [NickH]
LeeF: RDF working group resolves to remove untyped literal without an language tag now gets parsed as literal of type XSD:String
14:36:39 [LeeF]
paul's mail: http://lists.w3.org/Archives/Public/public-rdf-dawg/2011AprJun/0399.html
14:36:56 [NickH]
LeeF: the options that we have been discussing are
14:37:12 [NickH]
LeeF: 1) do nothing
14:38:35 [NickH]
LeeF: 3) Wait until the RDF working group has also resolved what should be done about pain literals with language tags
14:39:25 [NickH]
4) come up with a wording that says 'the is what SPARQL 1.1 does today, RDF-WG is working on something and when that is sorted out interpret the SPARQL 1.1 specification differently'
14:39:52 [NickH]
LeeF: Is making this changes significant enough to require a second last-call?
14:40:12 [AndyS]
LeeF: 2) Incorporate the current RDF WG resolution on simple literals, and nothing else.
14:40:13 [NickH]
options 2 and 3 would require a othing else. [1
14:40:37 [SteveH]
q+
14:40:38 [NickH]
LeeF: unsure if 4 will require a second last-call
14:40:48 [LeeF]
ack SteveH
14:41:28 [NickH]
SteveH: if we did 1 and added a note saying that the RDF-WG is currently investigating some issues, would it require a second last-call?
14:44:00 [AndyS]
INSERT DATA { :x :p "foo" . :x :p "foo"^^xsd:string } --> COUNT --> changes from 2 to 1
14:45:29 [AndyS]
... which is not an inference effect
14:46:06 [SteveH]
SteveH: I feel that the lowest risk way worward is to do 1) but with an informative note saying what we think will happen
14:48:09 [NickH]
AndyS: the real problem I see is the miss-alignment between the SPARQL-WG and the RDF-WG
14:48:22 [NickH]
AndyS: by the time that they get to rec, it will be rather late
14:48:26 [SteveH]
yes, we have to do lang tags as well
14:48:45 [NickH]
SPARQL 1.1.1?
14:49:52 [LeeF]
zakim, who's on the phone?
14:49:52 [Zakim]
On the phone I see kasei, cbuilara, Olivier, LeeF, NickH, AndyS, chimezie (muted), SteveH, MattPerry
14:49:53 [NickH]
I think we can't wait
14:50:06 [NickH]
but should make a note about it
14:50:20 [LeeF]
So, NickH, you're happy with Steve's suggested approach?
14:50:45 [NickH]
yes
14:51:32 [LeeF]
ACTION: Lee to talk to Sandro and Eric about the informative note approach to aligning SPARQL with the RDF literal changes
14:51:33 [trackbot]
Created ACTION-485 - Talk to Sandro and Eric about the informative note approach to aligning SPARQL with the RDF literal changes [on Lee Feigenbaum - due 2011-07-05].
14:53:36 [kasei]
service tests 2, 3, and 6 list missing data files in the manifest
14:53:53 [kasei]
service tests 4 and 7 are, I believe, wrong.
14:54:42 [LeeF]
qt:data <data02.ttl> ;
14:54:57 [kasei]
I'll send an email with the bigger problems with tests 4 and 7
14:55:34 [AndyS]
ericP (tests for federation?)
14:57:12 [NickH]
nopem sorry
14:57:18 [NickH]
but happy to learn!
14:57:25 [NickH]
ok, thanks
14:57:31 [Zakim]
-SteveH
14:57:32 [Zakim]
-LeeF
14:57:32 [Zakim]
-chimezie
14:57:34 [Zakim]
-NickH
14:57:35 [Zakim]
-AndyS
14:57:37 [Zakim]
-cbuilara
14:57:37 [Zakim]
-Olivier
14:57:40 [Zakim]
-MattPerry
14:57:48 [Zakim]
-kasei
14:57:49 [Zakim]
SW_(SPARQL)10:00AM has ended
14:57:50 [Zakim]
Attendees were kasei, cbuilara, LeeF, NickH, SteveH, Olivier, AndyS, chimezie, MattPerry
14:58:28 [LeeF]
SteveH - I'm happy with the JB-6 response
14:58:56 [SteveH]
thanks LeeF
14:59:10 [LeeF]
NickH, there are some instructions on how to use CommonScribe (the tool we use to generate the minutes) here: http://lists.w3.org/Archives/Public/public-rdf-wg/2011May/0298.html
14:59:16 [LeeF]
why don't you give it a shot and let me know if you have any issues?
14:59:24 [NickH]
LeeF: sure
15:01:18 [NickH]
how magical!
15:05:37 [NickH]
Done: http://www.w3.org/2009/sparql/meeting/2011-06-28
15:24:36 [LeeF]
awesome, thanks, NickH!
15:55:44 [SteveH]
AndyS, the text I added to JB-6 was: "We will add some text to the document making it clear that any function/operator which returns a type error may be extended by implementations, this has become less clear since the 1.0 version." — does that seem OK?
16:52:00 [Zakim]
Zakim has left #sparql
17:29:16 [AndyS]
sure - there are all sorts of traps in extending SUM and I don't see a necessary reason why it has to be done that way so (LC!) [I think casting is what is really needed]
19:03:39 [AndyS]
AndyS has joined #sparql
20:46:13 [SteveH]
SteveH has joined #sparql
20:46:25 [SteveH_]
SteveH_ has joined #sparql