IRC log of dawg on 2004-09-28

Timestamps are in UTC.

14:26:27 [RRSAgent]
RRSAgent has joined #dawg
14:26:33 [Zakim]
14:26:36 [Zakim]
14:27:01 [howardk]
howardk has joined #dawg
14:27:09 [Zakim]
14:27:12 [AndyS]
zakin, ??P25 is AndyS
14:27:18 [howardk]
morning all
14:27:20 [AndyS]
zakim, ??P25 is AndyS
14:27:20 [Zakim]
+AndyS; got it
14:28:13 [AndyS]
zakim, please dial ericP-617
14:28:13 [Zakim]
ok, AndyS; the call is being made
14:28:14 [Zakim]
14:28:40 [Zakim]
14:28:58 [Yoshio]
Good evening
14:29:39 [SimonR]
DanC: Possibly trouble -- I can't seem to reach Zakim. You might have to find a backup backup scribe.
14:29:48 [DanC]
14:29:58 [Zakim]
14:30:00 [DanC]
Zakim, pick a scribe
14:30:00 [Zakim]
Not knowing who is chairing or who scribed recently, I propose Yoshio
14:30:08 [SteveH]
Zakim, +??P28 is SteveH
14:30:08 [Zakim]
sorry, SteveH, I do not recognize a party named '+??P28'
14:30:10 [Yoshio]
14:30:18 [DanC]
Zakim, pick a scribe
14:30:18 [Zakim]
Not knowing who is chairing or who scribed recently, I propose Kendall_Clark
14:30:25 [DanC]
Zakim, pick a scribe
14:30:25 [Zakim]
Not knowing who is chairing or who scribed recently, I propose DaveB
14:30:27 [SteveH]
Zakim, ??P28 is SteveH
14:30:27 [Zakim]
+SteveH; got it
14:31:04 [Zakim]
14:31:06 [DanC]
DanC has changed the topic to: RDF Data Access 28 Sep. chair: DanC; scribe: DaveB
14:31:13 [DanC]
Zakim, next agendum
14:31:13 [Zakim]
agendum 1. "Convene, take roll, review record and agenda" taken up [from DanC]
14:31:48 [Zakim]
14:31:54 [DanC]
Zakim, who's on the phone?
14:31:54 [Zakim]
On the phone I see Yoshio, DaveB, Kendall_Clark, DanC, HiroyuiS, AndyS, EricP, HowardK, SteveH, PatH, [ASemantics]
14:31:57 [DaveB]
did JanneS volunteer scribing for next mtg?
14:32:04 [DaveB]
14:32:08 [patH]
patH has joined #dawg
14:32:09 [DaveB]
roll call
14:32:29 [DaveB]
as above oOn the phone, correct spelling
14:32:33 [Yoshio]
14:32:41 [DaveB]
regrets tomA, JanneS
14:32:56 [DaveB]
f2f3 meeting record
14:33:18 [DaveB]
DanC - copy of AndyS live editing from f2f3 not avail, is that OK?
14:34:07 [DaveB]
AndyS to send notes to email, or point at them if already sent
14:34:36 [DanC] + ammendment to come
14:34:43 [DanC]
14:35:00 [DaveB]
14:35:07 [DaveB]
5 Oct, EricP to chair
14:35:19 [DaveB]
JanneS scribe
14:35:47 [Yoshio]
possible regrets (moving next day)
14:35:51 [DanC]
Zakim, next agendum
14:35:51 [Zakim]
agendum 2. "Introduction and abstract (name, marketing)" taken up [from DanC]
14:36:11 [DanC]
14:37:22 [Zakim]
14:37:33 [Yoshio]
14:37:45 [SimonR]
(Yay, I got through on the phone!)
14:38:06 [DaveB]
Zakim, ??P30 is SimonR
14:38:06 [Zakim]
+SimonR; got it
14:38:18 [DaveB]
DanC - propose to edit in SPARQL for now, but not close the issue
14:38:56 [DaveB]
DaveB - and remove the other choices I hope
14:39:13 [DaveB]
ACTION AndyS change name to SPARQL, keeping issue open in rq23
14:39:16 [DanC]
editor is inclined to accept the feedback in the direction of SPARQL
14:39:43 [DaveB]
KendallC - words for abstract
14:39:51 [DaveB]
that action DONE
14:40:14 [DanC]
14:40:15 [Zakim]
14:40:27 [DanC]
Zakim, ??P32 in JosD
14:40:27 [Zakim]
I don't understand '??P32 in JosD', DanC
14:40:45 [DanC]
Zakim, next agendum
14:40:45 [Zakim]
agendum 3. "issue: SOURCE" taken up [from DanC]
14:40:58 [JosD]
JosD has joined #dawg
14:41:49 [Zakim]
14:42:35 [Zakim]
14:43:05 [DaveB]
re section9 of rq23
14:43:16 [DaveB]
editors inclined to publish as is
14:43:29 [DanC]
ACTION DaveB: Update the source section 9, add
14:43:29 [DanC]
more formal links, update the examples, try to think about extra
14:43:29 [DanC]
constraints as EricP proposed (SOURCE ?s and ?s onlin in SELECT). Look
14:43:29 [DanC]
at various ppl's source test cases.
14:44:35 [DanC]
ACTION SteveH: test cases (10 or so) re: current SOURCE design. in progress
14:44:44 [DanC]
ACTION AlbertoR: real test cases re: current SOURCE design. in progress
14:44:55 [kendall]
zakim, mute me
14:44:55 [Zakim]
Kendall_Clark should now be muted
14:45:10 [Yoshio]
* Yoshio ducked
14:45:20 [AndyS]
14:45:40 [kendall]
zakim, unmute me
14:45:40 [Zakim]
Kendall_Clark should no longer be muted
14:45:49 [DanC]
Andy's ACTION done
14:45:54 [DaveB]
re 0541 helps JosD to distinguish 2 designs
14:46:58 [DaveB]
AndyS re 0541
14:47:11 [DaveB]
Q2 (2a) has 2 row results
14:47:52 [DaveB]
in q2b there is no implicit distinct, no request for ?src, so get 1 result
14:48:33 [Yoshio]
Hmm, not feel natural with two lines
14:50:38 [ericP]
defns only
14:52:45 [DaveB]
discussion of duplicates in results, costs
14:52:58 [DaveB]
and refernece to Query Solution that isn't clear if it's a set of (set of bindings)
14:53:36 [DaveB]
PatH suggests being agnositic on duplicates, ref to systems that do resulting counting
14:54:46 [DaveB]
distinct will always give 1 right answer
14:54:51 [DaveB]
DISTINCT not in syntax yet
14:55:28 [kendall]
i've been thinking that limit might be a protocol bit
14:55:34 [Yoshio]
why so minmalistic? it would be hard to use with dupicates... at least unnatural...
14:56:17 [ericP]
yoshio, it has a cost on streaming
14:56:26 [AlbertoR]
14:56:32 [AndyS]
Can be useful to see that "Andy" occurs twice when projecting out in SELECT
14:56:43 [DanC]
alberto, tests based on ?
14:56:50 [ericP]
yoshio, you end up with all of your results in a hash table that is consulted before each new tuple gets streamed
14:57:00 [AndyS]
It also costs in terms of memory (need to keep track of previous things returned)
14:57:41 [Yoshio]
mmm, things are more complicated than I thought... but...
14:57:44 [SimonR]
ericP: Actually, in order for us to do streaming joins we have to make sure our intermediate results have a sort ordering. So a side-effect of streaming for us is that distinct is actually fairly easy....
14:58:16 [DaveB]
continue alberto's action, ref to 0327
14:58:32 [Yoshio]
how about allowing duplicates only when streaming
14:59:31 [DanC]
Zakim, next agendum
14:59:31 [Zakim]
agendum 4. "Toward 1st spec WD" taken up [from DanC]
15:00:11 [DaveB]
action dirk continued
15:00:50 [SteveH]
dont qnames have to have a :?
15:00:57 [DaveB]
action AndyS DONE
15:01:25 [DaveB]
SteveH: yeah. but in some n3's it is optional
15:02:07 [SteveH]
DaveB, how is that qualified?
15:02:20 [SteveH]
it would just be a URI wouldn't it?
15:02:48 [DaveB]
AndyS hoping for draft by 1 Oct/Friday
15:02:52 [DaveB]
15:03:15 [AndyS]
yes - I *think* we can could drop the ?/$ but hard for some lexers (I'm told)
15:03:39 [DaveB]
DaveB asks for 2 working days
15:04:01 [AlbertoR]
AndyS: talked with DW right now (not on IRC either) - he will make his action later when he has time - then we can discuss more about ? vs. $
15:04:12 [kendall]
fwiw, i find queries w/out var markers harder to read than ones with var markers.
15:04:15 [AndyS]
And it looks all the same (qnames / variables)
15:04:35 [DaveB]
ACTION HowardK: review updated spec (appearing 1 Oct) for 5th Oct
15:04:44 [SteveH]
I do think the ?'s help visually, but I see Alberto's concerns
15:05:01 [DanC]
Zakim, next agendum
15:05:01 [Zakim]
agendum 5. "Web Service Constraints etc." taken up [from DanC]
15:05:11 [DaveB]
(other review actions SteveH, DaveB continued)
15:05:31 [DaveB]
KendallC action DONE
15:05:43 [DaveB]
15:05:56 [Yoshio]
* re: ? and $ what has become of the problem when bookmarking?
15:06:19 [DaveB]
Yoshio: it makes little difference, you have to URI escape lots of characters anyway, for safety
15:06:48 [Yoshio]
DaveB: I see
15:06:52 [DaveB]
DanC - would be nice to have a UC re rdf query and web service constraints
15:07:34 [DaveB]
ACTION EricP: draft UC on overlap between RDF query and web service constraints
15:07:47 [DaveB]
... re WS-Policy
15:08:10 [DaveB]
also see OWL-S, lots of other semantic web services work
15:08:47 [DaveB]
possibly for a new UC draft 5th Oct
15:08:50 [DaveB]
15:09:09 [DanC]
Zakim, next agendum
15:09:09 [Zakim]
agendum 6. "Toward updated Use Cases & Requirements WD" taken up [from DanC]
15:09:39 [AlbertoR]
Yoshio: no is more about client side DBC API having problems dealing with '?' e.g. queries run through M$ Excel using an ODBC - you need wierd escaping of ? sings - Dirk will (hopefully) email more abou this later
15:09:46 [DanC] $Revision: 1.136 $ of $Date: 2004/08/31 13:49:06 $
15:09:54 [DaveB]
KendallC has UC to add identity management, DI, above query/WS one, (danbri one)
15:09:59 [DaveB]
action continued
15:10:18 [DaveB]
SteveH continued
15:11:15 [DaveB]
ACTION PatH: Review UC&R draft appearing around Oct1, for Oct5 telcon
15:12:26 [DanC]
Zakim, next agendum
15:12:26 [Zakim]
agendum 7. "ftf meeting schedule" taken up [from DanC]
15:13:01 [DaveB]
SteveH can host, haven't written the email
15:13:17 [DaveB]
southampton offers to host
15:13:18 [DaveB]
action done
15:13:22 [ericP]
if we're going to be cold, let's go all the way
15:13:34 [DanC]
Zakim, who's on the phone?
15:13:34 [Zakim]
On the phone I see Yoshio, DaveB, DanC, HiroyuiS, AndyS, EricP, HowardK, SteveH, PatH, [ASemantics], SimonR, ??P32, Kendall_Clark
15:13:49 [Yoshio]
Cool : )
15:14:09 [DaveB]
Yoshio pref Helsinki
15:14:30 [DaveB]
DaveB Southampton, then WDC
15:14:33 [DaveB]
DanC no pref
15:14:51 [DaveB]
Hiroyuki pref Helsinki
15:15:01 [DaveB]
AndyS Southampton, Helsink, WDC that order
15:15:03 [Yoshio]
* can we see aurora there?
15:15:18 [DaveB]
EricP Helsinki, Southampton, WDC (but pref Helsinki!)
15:15:27 [DaveB]
HowardK WDC (if !Bush) else Helsinki
15:15:45 [DaveB]
SteveH Southampton Helsinki WDC
15:15:55 [DaveB]
PatH WDC Southampton Helsinki
15:16:03 [DaveB]
Alberto Helsinki Southampton
15:16:21 [DaveB]
SimonR WDC or others no pref
15:16:26 [DaveB]
Kendall Helsinki
15:16:31 [DanC]
Zakim, ??P32 is JosD
15:16:31 [Zakim]
+JosD; got it
15:16:32 [DaveB]
15:17:16 [DaveB]
Helsinki would be regrets - JosD, PatH
15:17:53 [ericP]
it should be easy to eliminate
15:18:13 [AndyS]
Richard Branson is arranging it.
15:18:19 [kendall]
15:19:08 [DanC]
January 2005
15:19:08 [DanC]
Su Mo Tu We Th Fr Sa
15:19:09 [DanC]
15:19:11 [DanC]
2 3 4 5 6 7 8
15:19:13 [DanC]
9 10 11 12 13 14 15
15:19:15 [DanC]
16 17 18 19 20 21 22
15:19:17 [DanC]
23 24 25 26 27 28 29
15:19:19 [DanC]
30 31
15:21:19 [DaveB]
PROPOSED: 19-20 January 2005 Espo, Finland hosted by Profium
15:21:41 [DaveB]
15:21:43 [kendall]
yay vodka!
15:21:56 [ericP]
and saunas
15:21:58 [AlbertoR]
sauna too :)
15:22:13 [kendall]
and birch-leaved self-flagellation ^H^H^H^H oops :>
15:22:13 [ericP]
and hacheting holes in the ice so we can swim
15:24:24 [kendall]
15:24:34 [DanC]
28 February - 4 March 2005
15:24:34 [DanC]
W3C Technical plenary (per W3C member events calendar)
15:24:47 [DaveB]
ACTION EricP: find logistics re F2F5 at tech plenary in Boston March 28Feb-4Mar (2 days)
15:25:04 [DaveB]
some pref to Mon/Tue for f2f5
15:25:23 [kendall]
15:25:40 [DanC]
Zakim, close this agendum
15:25:40 [Zakim]
agendum 7 closed
15:25:41 [Zakim]
I see 2 items remaining on the agenda; the next one is
15:25:42 [Zakim]
8. Issues [from DanC]
15:25:51 [AndyS]
+1 to Mon/Tues
15:25:52 [DanC]
15:26:28 [DanC]
Zakim, next agendum
15:26:28 [Zakim]
agendum 8. "Issues" taken up [from DanC]
15:27:16 [DaveB]
all item 8 actions state DONE/not done as in agenda
15:27:19 [DanC]
Zakim, next agendum
15:27:19 [Zakim]
agendum 8 was just opened, DanC
15:27:35 [DaveB]
item 9
15:27:48 [DaveB]
kendallC walking tour action continued
15:27:59 [DaveB]
kendallC deadline is F2F4
15:28:18 [DaveB]
KendallC CC interface, attempt this week or next. continued
15:28:26 [DaveB]
KendallC xml format - done
15:28:38 [DaveB]
15:28:50 [DaveB]
DaveB format done
15:28:59 [DaveB]
15:29:07 [DaveB]
EricP action done
15:29:36 [DaveB]
KendalLC - working on protocol doc not complete yet
15:31:19 [ericP]
I'm really divided on the value of being validated by XML-Schema
15:31:25 [ericP]
or DTD
15:32:09 [AlbertoR]
15:32:13 [AlbertoR]
15:32:28 [ericP]
q+ to ask if dave tried to shorhorm either into an RDF-compatible form
15:33:12 [DanC]
ack AlbertoR
15:33:19 [AndyS]
15:33:30 [kendall]
q+ to ask about sparql var names fitting into xml names
15:33:35 [DaveB]
more explanation needd - xslt, xpath - add an example
15:34:58 [DaveB]
bnodes in results
15:35:02 [DaveB]
issues with that
15:35:09 [kendall]
q+ to ask what <hpage></hpage> means
15:35:18 [DaveB]
PatH - does it allow future queries use that term?
15:35:44 [DanC]
ack ericp
15:35:44 [Zakim]
ericP, you wanted to ask if dave tried to shorhorm either into an RDF-compatible form
15:35:52 [SteveH]
kendall, I think <foo/> means foo=NULL
15:36:12 [DaveB]
yeah, I didn't actually form the data & queries, sorry
15:36:14 [kendall]
15:36:17 [DanC]
then how to encode foo="", steveH?
15:36:21 [DaveB]
I needed something with datatypes
15:36:24 [SimonR]
SteveH: How do you represent an empty plain literal, then?
15:36:24 [kendall]
not sure what i think about that
15:36:31 [AlbertoR]
DanC - right
15:36:47 [SteveH]
DanC, agreed, I was guessing, no saying I liked it
15:36:49 [kendall]
example with xml literal would be good, IMO
15:37:19 [DaveB]
ah yeah; it'll be horribly entity encoded stuff
15:37:36 [kendall]
yep, but be good to show it, fwiw
15:37:46 [AlbertoR]
DaveB syntax is cool - much flatter than orginal - just wanna be sure it has got all feature we need
15:38:03 [SteveH]
I like result1.xml, modulo NULL too
15:38:36 [DanC]
ack andys
15:39:55 [DanC]
ack kendall
15:39:55 [Zakim]
kendall, you wanted to ask about sparql var names fitting into xml names and to ask what <hpage></hpage> means
15:40:31 [DaveB]
explict it has no value, in xml syntax
15:40:43 [DaveB]
... is ok says AndyS
15:41:52 [DaveB]
constraining variable names to fit in xml
15:42:01 [DaveB]
... matching xml name production or subset
15:42:15 [DaveB]
AndyS in grammar, variable names are XML NCName
15:43:06 [AndyS]
What about XML literals with same tag name?
15:43:10 [DaveB]
PatH suggests introducing 'var ?x'
15:43:19 [DaveB]
er, var x. then withdraws
15:43:27 [DanC]
I think he meant <var name="...">..</var>
15:44:17 [ericP]
15:44:37 [DaveB]
adv to schema with closed content model
15:44:49 [DanC]
15:44:52 [DaveB]
could be in editing with schemas
15:45:02 [DaveB]
Alberto - or generating records
15:45:03 [AndyS]
q+ to ask what we said about XMLliterals
15:45:10 [kendall]
yes, easier to write this stuff by hand in smart xml editors w/ a schema
15:45:19 [kendall]
but i don't much much stock in that use case
15:45:38 [DanC]
ack danc
15:46:29 [DanC]
ack andy
15:46:29 [Zakim]
AndyS, you wanted to ask what we said about XMLliterals
15:46:30 [DaveB]
hmm, could have <name xsi:type="xsd:anyURI">...</name>
15:47:07 [DanC]
hmm... xpath expression that accidently matches...
15:47:19 [AlbertoR]
15:47:24 [DaveB]
worst case could be getting back a query in the result
15:47:39 [DaveB]
suggestion to use <[!CDATA, or whatever the syntax is
15:48:00 [kendall]
zakim, mute me
15:48:00 [Zakim]
Kendall_Clark should now be muted
15:48:28 [kendall]
zakim, unmute me
15:48:28 [Zakim]
Kendall_Clark should no longer be muted
15:49:18 [kendall]
zakim, mute me
15:49:18 [Zakim]
Kendall_Clark should now be muted
15:50:28 [kendall]
zakim, unmute me
15:50:28 [Zakim]
Kendall_Clark should no longer be muted
15:50:44 [DaveB]
two results formats - var bindings, rdf graphs
15:50:52 [DaveB]
the latter we have rdf/xml as rec format, also ntriples,...
15:51:01 [AndyS]
+1 to returning RDF graphs as well as "XML for XML tools"
15:51:19 [DaveB]
re the protocol design, KendallC has considered this two result formats.
15:51:33 [DaveB]
DanC asks for examples of these cases
15:52:38 [DaveB]
possibility of SELECT returning var binidngs (xml) and also returning a subgraph
15:52:48 [DaveB]
... in rdf/xml or whatever
15:54:11 [DaveB]
DanC: the rdf graph result for variable bindings might either look like a) the query graph or b) an encoding of the variable bindings, like the test resutl format
15:54:18 [kendall]
15:55:00 [DaveB]
CONSTRUCT in the current doc present but not up-to-date linked with earlier sections
15:55:13 [DaveB]
KendallC - third kind is the ASK result boolean
15:55:14 [DanC]
(I think we're already pretty high-risk, so I encourage you to go for it, AndyS)
15:55:36 [AndyS]
15:55:41 [AlbertoR]
q- I will take question about result1.xml on list later
15:55:49 [AlbertoR]
15:56:16 [DaveB]
DaveB - need a format for ASK result without a protocol
15:56:38 [DanC]
regrets next week from JosD
15:56:46 [DanC]
15:56:51 [Zakim]
15:56:53 [Zakim]
15:56:54 [Zakim]
15:56:55 [Zakim]
15:56:58 [Yoshio]
15:57:00 [Zakim]
15:57:01 [Zakim]
15:57:03 [Zakim]
15:57:04 [Zakim]
15:57:05 [Zakim]
15:57:06 [Zakim]
15:57:07 [Zakim]
15:57:24 [AlbertoR]
AlbertoR has left #dawg
15:57:35 [HiroyukiS]
15:57:40 [DanC]
hm.... ASK result format: <aye/> vs. <nay/>. 1/2 ;-)
15:58:20 [DaveB]
15:58:33 [AndyS]
+1 (=true)
15:58:43 [DaveB]
how about +1, -1 or 0 (dunno)
15:59:27 [DaveB]
IIRC OWL tests had a single char answer for something
16:07:03 [DanC]
Zakim, who is on the phone?
16:07:03 [Zakim]
On the phone I see AndyS, EricP
16:07:15 [RRSAgent]
16:14:59 [Zakim]
16:15:00 [Zakim]
16:15:01 [Zakim]
SW_DAWG()10:30AM has ended
16:15:02 [Zakim]
Attendees were Yoshio, DaveB, Kendall_Clark, DanC, AndyS, HiroyuiS, EricP, HowardK, SteveH, PatH, [ASemantics], SimonR, JosD