IRC log of dawg on 2005-05-03

Timestamps are in UTC.

14:29:09 [RRSAgent]
RRSAgent has joined #dawg
14:29:09 [RRSAgent]
logging to
14:29:35 [Yoshio]
14:29:39 [DaveB]
Zakim, who's on the phone?
14:29:39 [Zakim]
On the phone I see ??P27, Yoshio
14:29:42 [Zakim]
14:29:45 [DaveB]
Zakim, ??P27 is DaveB
14:29:45 [Zakim]
+DaveB; got it
14:29:48 [Zakim]
14:29:49 [howardk]
howardk has joined #dawg
14:29:50 [Zakim]
14:29:52 [Yoshio]
zakim, mut me
14:29:52 [Zakim]
I don't understand 'mut me', Yoshio
14:29:55 [kendall]
(and i believe anyone can rename that abstract type in the GET URI serialization in their <service> section of a WSDL...)
14:29:55 [AndyS]
zakim, ??P30 is AndyS
14:29:55 [Zakim]
+AndyS; got it
14:30:11 [Zakim]
14:30:13 [DanC]
Meeting: RDF Data Access weekly
14:30:15 [kendall]
zakim, mute me
14:30:15 [Zakim]
Kendall_Clark should now be muted
14:30:46 [DanC]
Topic: Convene, take roll, review records and agenda
14:30:50 [DanC]
14:30:58 [AndyS]
Pref we suggest one comon one so simple clients don't have to lool in service description at all
14:31:01 [Zakim]
14:31:19 [Zakim]
14:31:30 [KevinW]
KevinW has joined #DAWG
14:31:33 [AndyS]
e.g. I write a support library and ship it to many people for use on many services
14:31:42 [kendall]
Andy: I'm fine w/ that. But "default-graph-uri" is the last thing we talked about on mail/phone-call... DanC's suggestion to the contrary notwithstanding IMO
14:31:49 [Zakim]
14:32:17 [AndyS]
Don't care about the name much - as long as it is stable!
14:32:22 [kendall]
14:32:30 [kendall]
yes, to scribing
14:32:31 [jannesaa]
jannesaa has joined #dawg
14:32:32 [AndyS]
DanC was jusy noting what people currently do, I think
14:32:37 [DanC]
Zakim, who is on the phone?
14:32:37 [Zakim]
On the phone I see DaveB, Yoshio (muted), Jeen_Broekstra, DanC, AndyS, Kendall_Clark (muted), ??P33, Kevin, HowardK
14:32:56 [DanC]
Zakim, ??P33 is JanneS
14:32:56 [Zakim]
+JanneS; got it
14:33:07 [DanC]
14:33:23 [DanC]
missing one of howard's actions
14:33:40 [DanC]
Janne was there
14:34:01 [kendall]
last week's minutes APPROVED, with Howard's action addition and with Janne attending (and with Jeen's misplaced "crap")...?
14:34:04 [JosD]
JosD has joined #dawg
14:34:48 [Zakim]
14:35:06 [DanC]
next meeting 17 May?
14:35:20 [Zakim]
Not knowing who is chairing or who scribed recently, I propose Kendall_Clark (muted)
14:35:23 [Zakim]
Not knowing who is chairing or who scribed recently, I propose AndyS
14:35:27 [kendall]
10 May -- no meeting because of WWW conflict
14:35:35 [kendall]
17 May -- meeting @ regular time, JanneS scribing
14:35:35 [Zakim]
14:35:38 [jannes]
jannes has joined #dawg
14:35:52 [kendall]
zakim, unmute me
14:35:53 [Zakim]
Kendall_Clark should no longer be muted
14:36:45 [kendall]
(adding rules workshop precis to agenda)
14:36:47 [DanC]
agenda + rules workshop notes
14:36:58 [DanC]
Topic: issue fromUnionQuery
14:37:41 [kendall]
ericp: current design is a subset of the implemented algae expressivity
14:38:34 [kendall]
algae separates databases from provenance
14:39:21 [kendall]
eric's action continued
14:39:37 [DanC]
. ACTION: KendallC to draft protocol section on specifying datasets
14:39:37 [DanC]
(fromUnionQuery) including WSDL types and HTTP examples
14:40:36 [DanC]
params: default-graph-uri named-graph-uri
14:41:28 [DanC]
<xs:element minOccurs="0" maxOccurs="1" ref="default-graph-uri"></xs:elemen>t
14:41:28 [DanC]
<xs:element minOccurs="0" maxOccurs="unbounded" ref="named-graph-uri"></xs:element>
14:41:44 [DanC]
(hmm... spaces between them or multiple occurences?)
14:42:59 [DaveB]
fwiw output-xslt-params isn't as important as having an output-xslt-mime-type
14:43:07 [DanC]
14:43:41 [AndyS]
Do we want to use "default" in rq23 so protocol and rq23 agree? We have tended to say "default", not "background" in discussions.
14:43:52 [DaveB]
I'm waiting for http defn, examples. nothing there yet
14:44:12 [DanC]
typo: </xs:elemen>t
14:44:38 [Yoshio]
q+ to ask what if no deafult-graph-uri is given
14:45:31 [kendall]
Andy: I'd like them to agree, so "default" works for me in rq23
14:45:32 [DaveB]
why are the complexType contents in a sequence? surely the contents is unordered?
14:46:50 [kendall]
DaveB: yr right. don't need seq. i'll think about that and fix.
14:47:29 [Yoshio]
zakim, unmute me
14:47:29 [Zakim]
Yoshio should no longer be muted
14:49:25 [kendall]
(ah, the xs:sequences are artifacts of my schema editor...)
14:50:42 [AndyS]
I'm happy
14:51:25 [DaveB]
q+ a minor point
14:51:47 [Yoshio]
please someone summary up the answer to my question?
14:52:11 [kendall]
Yoshio: the behavior of an implementation when there is no default-graph-uri specified is itself unspecified
14:52:46 [kendall]
"output-xslt-mime-type" what Dave wants...
14:53:05 [kendall]
i thought that would be one of the output params
14:54:01 [AndyS]
I only support client-applied XSLT (currently) - server adds stylesheet, client auto does the xform
14:56:08 [AndyS]
Dave - would XSLT on client be OK for your usage?
14:57:37 [AndyS]
(such XSLT may be client accessible only - firewalls)
14:57:48 [DaveB]
I was saying sparqlx is operation2 of the only interface
14:58:14 [DaveB]
like an xslt transform, it's likely/possible this isn't implemetned
14:58:22 [DaveB]
sparqlx status is ...?uncertain?
14:59:49 [ericP]
q+ to discuss consistent language vs. consistent services
15:00:26 [DaveB]
I also see the proto-wd has a bunch of identifiers for languages and (some) unregistered mime types or xml vbr etc. in the saddle section. hmm. should xml vbr give a mime type hint?
15:01:28 [AndyS]
I hope to implment when there is a spec (HTTP examples should be enough). SOAP later.
15:01:38 [DaveB]
in tuscany I used output-type
15:01:41 [DanC]
15:01:43 [DanC]
is what i heard
15:01:55 [DaveB]
and what I used at
15:01:56 [kendall]
output-xslt-mimetype work?
15:04:02 [kendall]
I agree that it makes things easier to teach users, but that's not a reason to *specify* it, IMO
15:04:46 [DaveB]
andy mentions he generates <?xml-stylesheet ... > optionally
15:04:59 [kendall]
+1 to eric's point
15:05:47 [DaveB]
can somebody type eric's point?
15:05:59 [DanC]
ericP: I'd rather leave XSLT outboard.
15:06:01 [kendall]
(sorry, hard to scribe and discuss at the same time)
15:06:31 [Zakim]
15:06:43 [kendall]
+q to make a point about query/queryx
15:07:26 [Zakim]
15:07:38 [AndyS]
zakim, ??P0 is AndyS
15:07:38 [Zakim]
+AndyS; got it
15:07:38 [ericP]
P0, is that you AndyS?
15:11:37 [kendall]
(well, fwiw, when it was *very* simple, i wasn't grudging. now that it's getting more complex, i'm more grudging. :)
15:12:13 [AndyS]
Tend to agree with KC
15:12:13 [DanC]
POLL: (0) no XSLT anywhere in SPARQL protocol, (1) ala Andy, a pointer that will come back in <?xml-stylesheet ?> for the client, (2) yes, standardize an interface where the server does XSLT, with mimetype and all
15:12:59 [DanC]
ack yoshio
15:14:16 [DanC]
2 0 0/1 - 0/1 0 - 0/1 0 0
15:14:24 [AndyS]
and Andy abstains
15:14:43 [DanC]
note Alberto and maybe morten does 2
15:15:01 [AndyS]
q+ to ask what is (non SWEB) practice
15:15:24 [DaveB]
3 impls of this vs 0 of wsdl :)
15:15:41 [kendall]
15:15:43 [ericP]
q- ericP
15:16:43 [jannes]
Dave, put +1 on a WSDL side, too. We've implemented such as thing about a year ago for our product (but not the one we're discussing here)
15:20:00 [DanC]
DaveB: how about a separate interface: query-and-transform
15:21:30 [DanC]
Topic: protocol publication candidate
15:21:38 [DanC]
. ACTION: kendall to uncommment saddle stuff, email to list
15:22:16 [DanC]
. ACTION: Yoshio to edit saddle descriptions provided by KendallC
15:24:05 [kendall]
I intended to try to have a conversation with Yoshio, on list, in response to his questions...
15:24:19 [kendall]
hoping others will join in :>
15:25:08 [kendall]
zakim, mute me
15:25:08 [Zakim]
Kendall_Clark should now be muted
15:25:37 [DanC]
ACTION: Yoshio to edit saddle descriptions provided by KendallC [Continues]
15:25:42 [AndyS]
15:25:45 [kendall]
yes, bijan owes the world answers (bijan and Jacek)
15:25:49 [kendall]
that's done
15:25:51 [DanC]
1.35 does ACTION: KendallC to incorporate EricP's privacy text, salting to taste
15:25:53 [kendall]
zakim, unmute me
15:25:53 [Zakim]
Kendall_Clark should no longer be muted
15:26:17 [DanC]
. ACTION: KendallC to ask WSD WG for clarification about HTTP binding of WSDL2, esp lists of URIs
15:26:32 [DanC]
15:26:41 [DanC]
. ACTION: DanC to generate SELECT, CONSTRUCT, ASK protocol traces
15:26:54 [kendall]
there was so much free time at the rules workshop! :>
15:27:24 [kendall]
sounds right to me
15:27:36 [DanC]
Topic: results format
15:27:37 [kendall]
critical path of protocol draft is still danc, eric, yoshio, kendall
15:27:54 [jeen]
q+ on status of rdf varbinding results format vs. xml varbinding format
15:27:57 [DanC]
EricP 1.29 looks ready to go
15:28:44 [kendall]
zakim, mute me
15:28:44 [Zakim]
Kendall_Clark should now be muted
15:29:24 [kendall]
zakim, unmute me
15:29:24 [Zakim]
Kendall_Clark should no longer be muted
15:29:29 [kendall]
(sorry, phone won't stop ringing)
15:30:18 [DaveB]
ref 2.3.3 of
15:31:06 [kendall]
sorry, dave :>
15:32:30 [kendall]
15:32:38 [kendall]
Jeen asks what the status of the RDF results format is
15:32:50 [kendall]
SteveH has an action to publish it as a Note at his leisure
15:32:53 [DaveB]
discussion lead to ericP pubbing rf1 and protocol at same time; rf1 waits for protocol
15:32:55 [kendall]
Which may never happen
15:33:12 [jannes]
60 mins up, I would need to run for tonight
15:33:44 [DanC]
Topic: issue: valueTesting
15:33:49 [kendall]
bye Janne
15:33:50 [Zakim]
15:33:56 [Yoshio]
15:33:57 [jannes]
jannes has left #dawg
15:34:07 [ericP]
15:34:08 [DanC]
. ACTION: EricP to propose a test re "x"^^mytype != "y"^^mytpe
15:34:10 [kendall]
EricP to propose a test case for valuetesting
15:34:27 [kendall]
wants the action to continue with progress noted
15:35:05 [DaveB]
I got it now
15:35:17 [DanC]
$Revision: 1.43 $
15:36:41 [DanC]
FILTER ?rot1 = ?rot2
15:36:43 [DaveB]
I wish I could see the results in that document
15:37:05 [DaveB]
the result is a 404 here
15:37:36 [DanC]
404 @
15:38:18 [DaveB]
there's a system error mesage for ne-fail
15:38:22 [DanC]
"cannot open 'data/ValueTesting/extendedType-ne-fail.rq' for printing: No such file or directory"
15:38:36 [kendall]
15:40:07 [DanC]
ACTION: DanC to send comment on non-use of
15:40:07 [DanC]
<> to schema comments list
15:41:09 [DanC]
Topic: issue: sort
15:41:39 [DaveB]
added something?
15:42:08 [ericP]
15:43:13 [DaveB]
they are not sorted results in those answers right now
15:43:23 [kendall]
(Some discussion of whether the test results are correct?)
15:43:33 [kendall]
(er, scratch that)
15:43:39 [DanC]
15:44:16 [DaveB]
I'd prefer if the vbr xml term was the same as in the
15:44:50 [jeen]
15:45:47 [DaveB]
vbr xml orders from 1
15:47:17 [kendall]
how do you spell that?
15:47:31 [AndyS]
Seems to be me
15:47:34 [kendall]
zakim, mute me
15:47:34 [Zakim]
Kendall_Clark should now be muted
15:47:42 [AndyS]
I'm muted here now
15:47:45 [DaveB]
discussion went to use rs:index and ordering from 1...
15:47:46 [kendall]
DC's a rough town!
15:48:03 [AndyS]
That email thread was small details only
15:48:15 [AndyS]
Have not tried todays test - rest were fine by me
15:48:20 [DaveB]
I found a typo in the sort/manifest.ttl - newline in a "literal"
15:48:25 [DanC]
andy, are you happy with the "more sorting test cases" thread state?
15:48:45 [AndyS]
Have not tried latest test but rest was OK
15:49:30 [kendall]
zakim, unmute me
15:49:30 [Zakim]
Kendall_Clark should no longer be muted
15:49:38 [AndyS]
Q1: non-determinate output
15:49:51 [DanC]
"ORDER BY" section of
15:50:12 [AndyS]
Hodl on ...
15:50:53 [Zakim]
15:51:03 [DanC]
DanC is looking for definitions text re order by
15:51:29 [AndyS]
sec 10.1
15:51:41 [Zakim]
15:51:53 [DaveB]
I just ran all the sort/*.rq tests - so the syntax is ok, but my sort checker runs sort(1) before comparing so that's not much use here!
15:51:56 [AndyS]
zakim, ??P0 is AndyS
15:51:56 [Zakim]
+AndyS; got it
15:52:59 [kendall]
15:53:33 [DanC]
PROPOSED: that 1.328 and @@TEST address the sort issue
15:54:53 [jeen]
15:55:11 [AndyS]
.me passed 8 of 8 after a CVS update - not looked at test 8 yet
15:55:28 [DanC]
PROPOSED: that 1.328 and the test including address the sort issue
15:56:16 [DanC]
poll... jeen... andy...
15:57:24 [DaveB]
there's no chance I'll do sorting in a week. maybe in 2 months
15:58:16 [kendall]
I will have looked at them carefully in a week.
15:59:11 [DanC]
DanC suggests that Andy do a "I hereby propose this for sort" msg to stir up any remaining comments
15:59:28 [DanC]
Topic: SPARQL QL edits
15:59:49 [DanC]
well, actually not.
16:00:22 [AndyS]
I declare victory on 7 action 1
16:00:25 [DanC]
ADJOURN. workshop chat after.
16:00:39 [Zakim]
16:00:45 [Zakim]
16:00:47 [KevinW]
KevinW has left #DAWG
16:00:48 [howardk]
bye all
16:00:58 [Zakim]
16:01:14 [Zakim]
16:01:43 [Zakim]
16:01:47 [Zakim]
16:02:50 [jeen]
jeen has left #dawg