IRC log of dawg on 2004-07-15

Timestamps are in UTC.

00:00:35 [ericP]
[discussion of 3.7 Limited Datatype Support]
00:01:44 [ericP]
robS: are we talking about strings or data type representations?
00:03:19 [ericP]
RESOLVED: accpet "4.8 Literal Search" as a DO
00:03:26 [ericP]
abstentions from ericP and robS
00:09:53 [Yoshio]
q+ to ask what do we mean "a subset" in requirement 3.7 Limited Datatype Support
00:10:21 [Yoshio]
s/do we mean/we mean/
00:10:33 [Yoshio]
s/mean/mean by/
00:11:14 [Yoshio]
00:13:47 [ericP]
00:20:34 [Yoshio]
q+ to ask if we don't have any problem when we apply path expression to RDF which the basic structure is not trees but meshes, even with loops
00:22:18 [Yoshio]
s/which the/whose/
00:22:55 [ericP]
[discussion of BRQL]
00:23:49 [ericP]
robS: we don't have to adopt BRQL. prefer that we start with XQuery.
00:24:07 [ericP]
... builds on existing specs...
00:24:35 [Yoshio]
by pathe expression I mean path-expression-based language like XQuery
00:25:29 [ericP]
kendall: starting with an XQuery engine isn't sufficient for everyone
00:25:40 [ericP]
... no implementation in Python
00:25:46 [kendall]
no, it isn't *possible* for everyone
00:26:27 [ericP]
robS: not asserting that it's simple
00:27:11 [ericP]
... can't say we've impelemented the RDF query layer
00:27:15 [ericP]
simon: we need a spec in front of us to work with
00:27:39 [ericP]
danc: [gives history, back to query workshop]
00:28:02 [ericP]
... sounds like an argument to pause for a while.
00:29:59 [ericP]
robS: going wtih an RDF-only query languge, NI will suffer
00:30:10 [ericP]
jeff: the semweb in general will suffer
00:30:52 [ericP]
simon: does that follow from not adopting XQuery, a language for querying hierarchical documents?
00:31:18 [ericP]
... the RDF stack is more similar
00:31:36 [ericP]
jeff: not saying that out of the box, it's a solution
00:32:17 [ericP]
kendall: it's not XQuery vs. start from scratch.
00:34:35 [ericP]
robS: XQuery took a long time because they were designing a language that would be easy to write
00:35:09 [ericP]
howard: i think it was the complexities of XQuery that made the spec so big.
00:35:57 [ericP]
robS: the RDF query langs have crappy presentation over good calculus
00:37:50 [ericP]
00:52:39 [Yoshio]
Hmm aren't we just looking at BRQL as ONE of candidates for the strawman?
00:52:59 [ericP]
00:53:16 [Yoshio]
I think we're going to look into XsRQL or N3QL as well
00:59:03 [Yoshio]
what is the definition of the word algebra? is it different from calculus?
00:59:20 [DanC_jam]
good question, Yoshio. I'm actually not certain.
00:59:41 [Yoshio]
(I don't know the definition of calculus, either, ;))
01:01:26 [Yoshio]
and what's the difference from data model?
01:02:05 [Yoshio]
As the data model, what we have is RDF (obvious?)
01:02:35 [Yoshio]
I think it's quite different from that of XML (a tree like document model)
01:05:20 [Yoshio]
* we have reached to a consensus that we are not likely to have a consensus easily
01:29:05 [AndyS]
AndyS has joined #dawg
04:32:07 [Zakim]
Zakim has left #dawg
05:36:18 [AndyS]
AndyS has joined #dawg
16:21:11 [RRSAgent]
RRSAgent has joined #dawg
16:22:40 [SimonR]
Considering the ebXML UC.
16:23:17 [DanC_jam]
6 Jun rev of ebXML
16:24:14 [DanC_jam]
Tue, 06 Jul 2004 11:26:20 -0400
16:26:56 [JosD]
JosD has joined #dawg
16:27:34 [howardk]
howardk has joined #dawg
16:31:18 [HiroyukiS]
HiroyukiS has joined #DAWG
16:32:50 [SimonR]
RobS: The drill-down mentioned in the UC is in the description logic domain rather than in basic RDF.
16:33:05 [DaveB]
I wonder if it's time to think about defining terms. A triple pattern is .... A graph pattern is a cnjunction of triple patterns. etc.
16:34:01 [AndyS]
16:34:12 [kendall]
seems relatively undangerous
16:34:48 [SimonR]
I think we've reached a point where a formal, abstract model of what a query is will be useful, and there's general agreement that this is the case.
16:43:13 [ericP]
agenda request: winter f2f schduling and spring tech plen -- ask if "the next 2-4 weeks" is sufficient notification to schedule our winter f2f.
16:43:47 [kendall]
heh, reschedule :>
16:43:58 [SimonR]
Issue with drill-down is that querying for what counts as a direct descendant in a hierarchy rather simply a descendant in a hierarchy. To do this would require being able to query to non-existence of any intermediate points in the hierarchy.
16:45:17 [SimonR]
(...rather than simply...)
16:47:05 [ericP] has been chacl'd
16:49:10 [DanC_jam]
regrep tc
16:49:50 [JosD]
RIM 2.5
16:57:12 [AndyS]
Discussing ::
17:03:13 [SimonR]
ebXML use case motivates interactive browsing; Kendall feels that 2.10 already implies interaction browsing, but some disagreement that this is clearly implied.
17:03:55 [SimonR]
ACTION: Kendall add ebXML UC
17:04:52 [SimonR]
ACTION: RobS write email to Farrukh
17:05:19 [DanC_jam]
1.5 Relationship to XQuery
17:09:48 [SimonR]
EricP: Describes a UC involving querying the Edgar database and integrating XML and RDF sources into the one query.
17:19:10 [DanC_jam]
... exploration of XQuery/RDF query integration scenarios: is loose coupling sufficient? is there motivation for anything more than string concatenation to make up queries, and XML to return results?
17:21:25 [ericP]
example FA query:
17:22:02 [DaveB-lap]
DaveB-lap has joined #dawg
17:30:10 [SimonR]
Break at 10:30 until 10:45.
17:46:47 [HiroyukiS]
HiroyukiS has joined #DAWG
17:50:49 [ericP]
18:00:43 [DanC_jam]
Zakim, this is dawg
18:00:43 [Zakim]
ok, DanC_jam; that matches SW_DAWG(f2f)2:00PM
18:00:52 [DanC_jam]
Zakim, who's on the phone?
18:00:52 [Zakim]
On the phone I see RobS
18:01:00 [DanC_jam]
Zakim, RobS is temporarily MeetingRoom
18:01:00 [Zakim]
+MeetingRoom; got it
18:02:01 [AndyS]
Resuming ...
18:02:42 [DaveB]
DaveB has joined #dawg
18:02:44 [howardk]
FYI: code for traversing a possibly cyclic foaf:knows network in XsRQL:
18:03:55 [Zakim]
+ +17907aaaa
18:04:06 [DaveB-lap]
that's me
18:04:12 [DanC_jam]
Zakim, aaaa is DaveB
18:04:12 [Zakim]
+DaveB; got it
18:09:53 [DanC_jam]
AndyS tweaks title page... author/editor stuff...
18:10:38 [Yoshio]
what's the URI?
18:10:49 [DaveB-lap]
18:11:17 [Yoshio]
thank you, dave
18:12:06 [DaveB-lap]
18:13:43 [DaveB-lap]
18:15:33 [Yoshio]
BBQ :)
18:19:15 [DanC_jam]
DanC gave editorial input on TOC and argued against SOURCE
18:19:36 [DanC_jam]
(slightly longer version is recorded by AndyS in the live copy)
18:25:00 [TomAdams]
TomAdams has joined #dawg
18:34:45 [DanC_jam]
ACTION: Jos: explain log:includes to inform the discussion of SOURCE (nee provenance)
18:35:15 [DaveB-lap]
the SOURCE thing could be an extension point
18:35:50 [DaveB-lap]
I kinda like the way that algae allows it (triple) {foo:bar} where they are extensions, annotations of the triple, kinda
18:36:18 [kendall]
triple facets, i keep calling them.
18:36:54 [DaveB-lap]
ptr to jos's test
18:36:59 [DanC_jam]
jos refers us to example 2 in
18:42:01 [DanC_jam]
DaveB: I like the the way that algae does it [optionals] too. [pointer, please?]
18:42:26 [DaveB-lap]
I was saying I like the way the result triples are built up
18:42:30 [DaveB-lap]
in algae, and here
18:42:46 [DanC_jam]
ACTION AndyS: explain DESCRIBE design implicit in BRQL spec
18:42:53 [DaveB-lap]
if it means lessspecial features in the language, great
18:43:20 [DanC_jam]
a pointer to the relevant algae docs would be nice, daveb.
18:43:27 [DaveB-lap]
18:44:14 [DaveB-lap]
algae2 result sets description:
18:51:18 [DanC_jam]
ACTION DanC: explain evolution of log:semantics/log:includes from uri-is-graph to uri-is-doc in cwm, to inform discussion of SOURCE
18:51:41 [DaveB-lap]
I'm willing to take an action to explain the two uses of redland contexts I see most.
18:52:02 [DanC_jam]
very well, dave. pls write the action
18:52:43 [DaveB-lap]
ACTION DaveB: explain the main uses seen for redland contexts with respect to the provenance
18:53:15 [DanC_jam]
(I just discovered .)
18:53:39 [DaveB-lap]
yeah, they use the 4th item as something else, but they recently changed. from bnode|uri to just allowing a uri
18:58:56 [Yoshio]
q+ to clarify if NOT operator can have multiple triples as operands
19:00:51 [Yoshio]
19:00:56 [DaveB-lap]
Yoshio: as I understand it, no
19:02:07 [JosD]
Re nested optionals see also test case in
19:04:05 [Yoshio]
q+ to ask the meaning of DISTINCT
19:04:53 [DaveB-lap]
Yoshio, as I recall we based it on SeRQL - duplicate triples are removed from the output graph
19:05:08 [DaveB-lap]
or duplicate binding resutls
19:05:09 [Yoshio]
Yes, but is it optional?
19:05:28 [DaveB-lap]
if not given, you may get duplicates
19:05:57 [Yoshio]
Hmm, I'm not sure what's the use of having duplicates
19:06:45 [DaveB-lap]
it's a performance thing partially
19:07:02 [DaveB-lap]
the server (or something) can filter it for the end user
19:07:29 [DaveB-lap]
see SeRQL manual for lots of this
19:07:39 [Yoshio]
Thank you, I'll check it
19:07:54 [Yoshio]
19:07:54 [AndyS]
19:08:03 [DanC_jam]
^ a test manifest
19:08:56 [DanC_jam]
DanC: what test coverage do we have?
19:09:09 [DanC_jam]
A: for RDQL, a test suite that some developers are happy with. For other BRQL features, a start.
19:10:17 [Yoshio]
BTW, where are ericP, Simon, Kendall and Rob?
19:11:01 [DanC_jam]
ACTION Jos: discuss test suite documentation and maintenance with Steve, EricP, AndyS, ...
19:13:56 [kendall]
Yoshio: I had to use the phone for a second. My boss called, wanted to know what was going on. :>
19:14:16 [DanC_jam]
Jos: in the case of bindings, results can just be a list
19:14:19 [Yoshio]
welcome back, kendall, nice to have you again :)
19:14:27 [DanC_jam]
AndyS: need the names too. [explains why...]
19:15:38 [DaveB-lap]
q+ agenda for after lunch? f2f page isn't updated
19:15:40 [Yoshio]
what do you mean by "need the names too", Andy, I think I missed your comment
19:16:21 [kendall]
Yos: thanks :>
19:16:42 [DaveB-lap]
q+ to ask about the agenda for after lunch? f2f page isn't updated for day2
19:17:15 [DanC_jam]
(my mind hasn't been updated either)
19:17:45 [DaveB-lap]
19:19:30 [Yoshio]
* After lunch? Dessert! :)
19:21:02 [kendall]
a cuppa!
19:31:22 [DanC_jam]
(discussion of editorial structure)
19:33:34 [Yoshio]
From my own experience, it was very good that I could find the Concepts and Abstract Syntax document for RDF as a separate document
19:35:01 [Yoshio]
It helps me to understand clearly that RDF/XML is just one way of serialization (wrong?)
19:35:28 [DaveB-lap]
we clearly needed that since peoplw ere confused seeing the XML as the RDF
19:36:22 [Yoshio]
Yes, one of the members of INTAP Semantic Web Committee is one of them
19:39:03 [DanC_jam]
agenda + strawman ql
19:39:09 [DanC_jam]
agenda + protocols
19:39:17 [DanC_jam]
agenda + bookmarking requirement/objective
19:44:37 [DanC_jam]
agenda + toward a 1 Aug requirements/use cases update. recruit reviewers
19:47:23 [TomAdams]
zakim, who is on the phone?
19:47:24 [Zakim]
On the phone I see MeetingRoom, DaveB
19:47:51 [Zakim]
19:47:52 [Zakim]
19:47:52 [Zakim]
SW_DAWG(f2f)2:00PM has ended
19:47:53 [Zakim]
Attendees were MeetingRoom, +17907aaaa, DaveB
19:51:24 [DanC_jam]
$Revision: 1.11 $ of $Date: 2004/07/15 19:51:01 $
19:51:25 [DanC_jam]
20:08:29 [DanC_jam]
Zakim, agenda?
20:08:29 [Zakim]
I see nothing on the agenda
21:34:27 [AndyS]
AndyS has joined #dawg
21:34:57 [kendall]
kendall has joined #dawg
21:36:21 [DanC_jam]
DanC_jam has joined #dawg
21:36:41 [Yoshio]
Yoshio has joined #dawg
21:38:22 [afs]
afs has joined #dawg
21:40:20 [afs_]
afs_ has joined #dawg
21:41:08 [ericP]
21:41:25 [SimonR]
SimonR has joined #dawg
21:47:35 [HiroyukiS]
HiroyukiS has joined #DAWG
21:47:52 [DanC_jam]
agenda + UC&R publication update, pending updates
21:48:04 [DanC_jam]
agenda + Strawman Syntax
21:48:09 [DanC_jam]
agenda + Protocols (GetData), etc.
21:49:11 [JosD_scribe]
=========== discussing UC&R proposed text changes
21:49:28 [JosD_scribe]
expected next publishing date aug 1
21:50:15 [JosD_scribe]
going document order and then new ones? no, go through Kendall's todo list
21:51:25 [JosD_scribe]
... walking through Kendall's list
21:53:42 [JosD_scribe]
... finished the list
21:53:56 [JosD_scribe]
=== bookmarkable queries
21:54:24 [JosD_scribe]
suggested text:
21:54:31 [DanC_jam]
on 3.8
21:56:26 [JosD_scribe]
DanC: hope to make that testable
21:57:09 [JosD_scribe]
no objections; astaining 1 (SimonR)
21:58:51 [JosD_scribe]
RESOLVED: adopt suggested text in as objective 4...
22:00:57 [JosD_scribe]
=== XQuery compatible surface syntax
22:01:30 [JosD_scribe]
EricP will project the 2 examples
22:03:04 [DanC_jam]
query syntax proposal
22:04:50 [JosD_scribe]
(takes some tome to boot the projector :))
22:05:01 [JosD_scribe]
22:06:02 [ericP]
22:11:22 [JosD_scribe]
22:17:27 [JosD_scribe]
quting from the message "Therefore, NI proposes that a new requirement be considered by this group: The query language shall have an XQuery compatible concrete language syntax"
22:17:41 [JosD_scribe]
22:18:49 [JosD_scribe]
RobS: "XQuery compatible concrete language syntax" is syntax plus more??
22:19:35 [JosD_scribe]
RobS: suggesting "srface syntax"
22:20:36 [JosD_scribe]
22:22:28 [JosD_scribe]
straw poll shows that support as requirement is 1 and about 4 as an objective
22:29:15 [JosD_scribe]
EricP: copy wording from charter - Eric now gives some background...
22:30:47 [JosD_scribe]
... produce a binding for XQuery
22:32:33 [JosD_scribe]
DanC: propsed req "syntax we choose should exploit people knowledge of SQL"
22:32:40 [DanC_jam]
22:32:47 [DanC_jam]
1 or 2 in support, per straw poll
22:34:06 [JosD_scribe]
22:34:24 [JosD_scribe]
22:41:52 [JosD_scribe]
after discussion no shift in positions
22:44:36 [AndyS]
. ACTION: SimonR Track the work of the WG to keep a current, discussable XQuery syntax document
22:46:10 [DanC_jam]
write a document discussing tradeoffs with adapting XQuery to do [...?]
22:46:55 [DanC_jam]
ACTION: write a document discussing tradeoffs with adapting XQuery as an RDF query language
22:47:15 [AndyS]
Action on SimonR
22:47:53 [JosD_scribe]
=== requirement on disjunction
22:48:35 [AndyS]
22:48:36 [Yoshio]
* what kind of hats are we wearing now?
22:48:57 [DanC_jam]
on disjuction...
22:50:15 [JosD_scribe]
stated as requirement, appeals to 4 to 5
22:50:40 [Yoshio]
appeals to Yoshio, too
22:52:51 [DanC_jam]
ACTION RobS: help kendal find supporting use cases for disjuction
22:53:01 [Yoshio]
q+ to ask if we take disjunction of matching pattern, will it reduce the number of the match?
22:57:08 [JosD_scribe]
Yoshio: ask if we take disjunction of matching pattern, will it reduce the number of the match?
22:57:27 [JosD_scribe]
DanC: it is Q2 V Q1
22:59:34 [AndyS]
q+ On last clause of req - unclear on ASK queries implication
22:59:51 [AndyS]
q+ to ask On last clause of req - unclear on ASK queries implication
23:00:37 [DaveB]
I'd like to know if there is complexity when optionals & disjunction are used together.
23:01:00 [DaveB]
well, I bet there is, but can it be explained
23:01:31 [AndyS]
Maybe optionals is actually the disjunction effect but done in a different way.
23:02:22 [JosD_scribe]
AndyS: On last clause of req - unclear on ASK queries implication
23:03:33 [JosD_scribe]
DanC proposes his test case of last night
23:04:46 [DanC_jam]
"The query language must include the capability to restrict matches on a
23:04:47 [DanC_jam]
queried graph based on a disjunction of graph patterns, at least one of
23:04:47 [DanC_jam]
which must be satisfied"
23:05:49 [ericP]
annotea use case:
23:05:54 [JosD_scribe]
RESOLVED disjunction as requirement
23:05:54 [ericP]
23:06:15 [JosD_scribe]
no objections, 1 abstention (JosD)
23:06:47 [Yoshio]
I'm happy with the disjunction being adopted as a requirement, but I fell more clarification in its wording
23:06:59 [Yoshio]
23:07:17 [Yoshio]
s/clarification/clarification needed/
23:07:59 [DanC_jam]
ACTION Kendal: draft revision, toward updating our public WD, delivery ~next wed.
23:08:03 [DanC_jam]
ACTION RobS: review it.
23:08:24 [DanC_jam]
ACTION 10= Kendall: draft revision, toward updating our public WD, delivery ~next wed.
23:09:00 [DanC_jam]
23:09:04 [Yoshio]
23:09:07 [DanC_jam]
Zakim, close agendum 1
23:09:07 [Zakim]
agendum 1 closed
23:09:08 [Zakim]
I see 2 items remaining on the agenda; the next one is
23:09:08 [Yoshio]
23:09:09 [Zakim]
2. Strawman Syntax [from DanC_jam]
23:09:11 [AndyS]
23:09:16 [JosD_scribe]
now we have a plan for publishing an updated UC&R document
23:09:47 [Yoshio]
q+ to ask if you feel some means to easily search for items in a collection
23:12:41 [DanC_jam]
. ACTION DanC: notify Semantic Web CG of risks around the "1.5 Relationship with XQuery" scope of our charter.
23:20:35 [DanC_jam]
ACTION DanC: notify Semantic Web CG of risks around the "1.5 Relationship with XQuery" scope of our charter.
23:21:01 [Yoshio]
23:21:15 [JosD_scribe]
============ strawman syntax
23:21:16 [DanC_jam] 1.11
23:23:03 [DanC_jam]
30 oct RDQL
23:23:54 [DaveB]
the above isn't public, you likely mean
23:24:38 [JosD_scribe]
who would prefer RDQL: 3
23:25:41 [JosD_scribe]
who would prefer BRQL over RDQL: about 6
23:26:52 [JosD_scribe]
advive from chair is that RDQL will bring us faster to REC
23:27:05 [JosD_scribe]
23:29:20 [JosD_scribe]
there are now issues raised in
23:29:54 [Yoshio]
23:31:08 [JosD_scribe]
critical mass for support of 1.11
23:32:54 [JosD_scribe]
RESOLVED: adopt 1.11 as strawman - objection NI - abstention HowardK
23:33:46 [DaveB]
congrats all
23:33:47 [DanC_jam]
NEW ISSUES: re CONSTRUCT: what happens when variable are not bound?
23:34:05 [DanC_jam]
NEW ISSUE: BARQL 1.11 does not support yes/no queries sufficiently
23:34:26 [DanC_jam]
23:34:28 [Yoshio]
q+ to ask if we're going to have other strawmen
23:35:06 [DanC_jam]
ack yoshio
23:35:06 [Zakim]
Yoshio, you wanted to ask if we're going to have other strawmen
23:35:18 [DanC_jam]
yes, we have an action on parallel XQuery design
23:36:31 [SimonR]
23:37:40 [DanC_jam]
Kendall: if requirements settle down, I might be willing to maintain the issues list
23:37:40 [JosD_scribe]
Kendall: make it easier to find the issues; Kendall is offering himself to maintain an issue list!!!
23:38:08 [JosD_scribe]
======== protocols
23:39:22 [kendall]
23:39:30 [DanC_jam]
# evalutate TAP's GetData? Dan Connolly (Sunday, 4 July)
23:40:30 [DanC_jam]
23:45:27 [kendall]
dan would like getdata folks to support what we do
23:45:44 [kendall]
seems unlikely since we may be more complex, and they've tried to maximize simplicity
23:47:02 [JosD]
Motivation for GetData
23:49:14 [DanC_jam]
cf TAP: A System for integrating Web Services into a Global Knowledge Base. R.V.Guha and Rob McCool
23:49:23 [DanC_jam]
23:55:06 [JosD]
Joseki test suite is Junit based; does graph isomorphism matching; ...