IRC log of dawg on 2004-04-01

Timestamps are in UTC.

15:29:38 [RRSAgent]
RRSAgent has joined #dawg
15:29:43 [DaveB]
DaveB has joined #dawg
15:29:43 [DanC]
Zakim, this will be RDFD
15:29:43 [Zakim]
ok, DanC; I see SW_DAWG()10:30AM scheduled to start in 1 minute
15:30:01 [AndyS]
At TOC 1.1
15:30:17 [DanC]
agenda + convene, take roll, review record, agenda
15:30:22 [DanC]
agenda + Use Cases
15:30:31 [DanC]
agenda + Requirements
15:30:38 [DanC]
agenda + Issues
15:30:44 [DanC]
agenda + Amsterdam meeting arrangements
15:31:02 [DanC]
15:32:11 [DanC]
Zakim, this will be RDFD
15:32:11 [Zakim]
ok, DanC; I see SW_DAWG()10:30AM scheduled to start 2 minutes ago
15:32:41 [DanC]
Zakim, this is DAWG
15:32:41 [Zakim]
ok, DanC; that matches SW_DAWG()10:30AM
15:32:45 [Zakim]
15:33:04 [Zakim]
15:33:22 [patH]
patH has joined #dawg
15:33:22 [eikeon]
There's a bit of an echo today.
15:33:30 [jannes]
jannes has joined #dawg
15:33:57 [Zakim]
15:34:14 [DanC]
Zakim, aacc is AlbertoG
15:34:14 [Zakim]
+AlbertoG; got it
15:34:21 [Zakim]
15:35:02 [DanC]
Zakim, pick a scribe
15:35:02 [Zakim]
Not knowing who is chairing or who scribed recently, I propose RobS
15:35:34 [Zakim]
15:35:56 [DaveB]
roll call
15:36:03 [DaveB]
alberto - asemantics
15:36:19 [DaveB]
andyS - hp
15:36:21 [DaveB]
kevinW - hp
15:36:27 [DaveB]
jean francois - inria
15:36:28 [KevinW]
KevinW has joined #dawg
15:36:40 [DaveB]
kendall , daniel - maryland
15:36:46 [DaveB]
rob - network inf
15:36:55 [dirkx]
dirkx has joined #dawg
15:36:58 [DaveB]
patricks - nokia
15:36:59 [AndyS]
zakim, ??P32 is AndyS
15:36:59 [Zakim]
+AndyS; got it
15:37:03 [DaveB]
janne - profium
15:37:11 [DaveB]
daveB - bristol
15:37:14 [DaveB]
pat hayes
15:37:19 [DaveB]
- invited expert
15:37:27 [DaveB]
katz - apologies
15:37:30 [DaveB]
15:37:32 [DaveB]
15:37:36 [DaveB]
-- end of roll call
15:37:36 [DanC]
Zakim, take up agendum 1
15:37:36 [Zakim]
agendum 1. "convene, take roll, review record, agenda" taken up [from DanC]
15:37:47 [DaveB]
s/apologies/regrets/ above for katz
15:37:53 [DaveB]
regrets for andyS next telcon
15:37:59 [DanC]
Zakim, pick a scribe
15:37:59 [Zakim]
Not knowing who is chairing or who scribed recently, I propose EricP
15:38:12 [DaveB]
scribe next telcon - ericP
15:38:29 [DaveB]
donm 8 april
15:38:40 [DaveB]
still looking for a co-chair
15:38:43 [DaveB]
15:30 UTC
15:39:26 [DaveB]
alberto - swbpd telcon 15:00 UTC, thursday - overlaps
15:39:42 [DaveB]
patHayes also on swbpd
15:39:48 [Zakim]
15:39:50 [DaveB]
danC - reviewing times after/at F2F
15:40:28 [DaveB]
minutes approved of last meeting
15:40:53 [DanC]
Zakim, next item
15:40:53 [Zakim]
agendum 2. "Use Cases" taken up [from DanC]
15:40:56 [DaveB]
agenda 0230.html
15:41:07 [kendall]
heavy breather is back :>
15:41:19 [DanC] a draft DAWG document outline
15:41:25 [ericP]
do we want to adderss action items?
15:41:42 [DaveB]
kendall outlines 0231 above
15:42:09 [DaveB]
KC - user stories I liked
15:42:09 [ericP]
someone is breahting into the microphone
15:42:31 [DaveB]
KC - motivating me, classifying by problem domain
15:42:51 [DaveB]
KC - AndyS also did a draft, more stuff with user storeies. going to work to try to merge them into 1 doc
15:43:33 [DaveB]
?not mentioned xqueyr
15:44:01 [thompsonbry]
thompsonbry has joined #dawg
15:44:01 [DaveB]
KC - an oversight, to add
15:44:21 [JosD]
JosD has joined #dawg
15:44:40 [DaveB]
an outline, generally seems ok
15:45:00 [Zakim]
15:45:03 [DaveB]
ACTION ericP, kendallC: sort out CVS access, put this on the home page
15:45:43 [DaveB]
DanC - this isn't the final doc, or even includes all the UCs
15:46:02 [DaveB]
KC - motivation for people who care about a UC to flesh it out, add a story - stop it being so thin
15:46:20 [DaveB]
note to scribe - aagfa people arrived? JosD & others?
15:47:05 [DaveB]
present +JosD - agfa
15:47:39 [DaveB]
PatrickS - would be nice to have a summary at the start to describe where it is going
15:48:01 [DaveB]
... what the purpose of the doc is, possible requirements
15:48:18 [DaveB]
... each UC have a 1 paragraph summary?
15:48:46 [DaveB]
KC - love to have UCs with a para of text, easier to read
15:48:56 [ericP]
15:49:26 [DanC]
Zakim, next item
15:49:27 [Zakim]
agendum 3. "Requirements" taken up [from DanC]
15:49:45 [thompsonbry]
bryan thompson also joined recently.
15:50:07 [thompsonbry]
like to see addressing and linking added to the set of related technologies and standards.
15:50:17 [DaveB]
ok, bryan - ack
15:50:48 [DaveB]
EricP discusses url above
15:50:53 [DaveB]
implemntation survey
15:51:03 [ericP]
15:51:15 [DaveB]
... section "Implentnation Survey"
15:51:29 [ericP]
15:51:32 [DaveB]
top part is an attempt of an abstract syntax
15:51:42 [DaveB]
anchors for language characteristics we could use
15:51:51 [DaveB]
to describe what the different existing languages do
15:51:56 [DaveB]
and what the UCs need/usr
15:51:57 [DaveB]
15:52:24 [DaveB]
happy to add items that aren't covered to this doc
15:52:36 [DaveB]
AndyS - a lot more than just queyr langs (QLs)
15:52:43 [DaveB]
... such as rules
15:52:53 [DaveB]
... and we need to be aware of that. Has to be query aspect.
15:53:10 [DaveB]
EricP - rules, a fair amount in the doc above
15:53:16 [kendall]
zakim, queue
15:53:16 [Zakim]
I don't understand 'queue', kendall
15:53:19 [kendall]
15:53:39 [DaveB]
EricP - ~70% query
15:54:10 [kendall]
zakim queue+ me
15:54:11 [DaveB]
ACTION EricP: Send mail to group about rdf-query-rules doc, proposing to using it an abstract syntax to describe requirements
15:54:29 [DaveB]
KC - don't see anything about rdf path languages
15:54:51 [DaveB]
KC - considering this, or out of scope?
15:55:04 [DaveB]
EricP - there are a couple there SeRQL, Xqueyr, ... -
15:55:28 [DaveB]
... abstract syn part doesn't really describe serialations, mostly about functions such as comparing rdf literals
15:55:52 [DaveB]
... might be tricky to add to this document.
15:56:18 [DaveB]
PatrickS - looks useful, I like it. But don't want to be constrained to things in this survey doc.
15:56:40 [DaveB]
EricP - will be responsive to adding new terminology as people need them
15:56:47 [kendall]
ericp: i had Versa in mind specifically, FWIW.
15:57:13 [DaveB]
PatrickS - just want WG to be clear this isn't scoping us
15:57:57 [DaveB]
Pat Hayes (PH) - tree structure / graph - presumption graphs are a tree structure?
15:58:07 [DaveB]
EricP - no, but could clarify with an example
15:58:16 [DaveB]
.. however some path languages don't express things that aren't trees
15:59:08 [DaveB]
PH - implementations?
15:59:15 [DaveB]
EricP - mostly stuff I found good docs for
15:59:48 [AndyS]
ericp: Could you add the "tell me about .." and Patricks URIQA and CBD
15:59:49 [alberto]
other list of RDF query languges (still maintained)
16:00:00 [DaveB]
16:00:16 [rob]
I'm the guy with my own QL...
16:00:21 [DaveB]
DanC - tell EricP if something you want, have, isn't there
16:00:45 [DaveB]
Jos deRoo (JD) - access aspect present?
16:00:47 [DanC]
no, dave, I said "if...". I guess if you misunderstood, others did too.
16:01:14 [DaveB]
16:01:20 [rob]
I must admit I do think we're getting ahead of ourselves.
16:01:42 [rob]
I think we've got a little more requirement-gathering before moving on to full query languages.
16:03:08 [DaveB]
AndyS - don't want requirements to displace work on UCs
16:03:23 [DaveB]
DanC - want to work on both
16:04:05 [rob]
Is it worth trying to get consensus on what the "tell me about" requirement/usecase means?
16:04:08 [DaveB]
.. UCs are there to 1) tell the world about this tech 2) for us so we know what we mean when we make requiremetns
16:05:13 [DaveB]
DanC - we are discusing things not UCs or requirements somewhat
16:05:31 [DaveB]
DanC - optional triples for example
16:05:41 [DaveB]
worth having in candidate reuqirements?
16:06:09 [DaveB]
support for optional triples as a candidate requirement - AndyS (proposer), Alberto, DaveB
16:06:30 [DaveB]
DanC - disjunction? use case of Rob (ref?)
16:07:08 [kendall]
that's 0184 in the archive
16:08:08 [thompsonbry]
graph update as candidate requirement ;) ?
16:08:08 [DaveB]
candidate requirement - DanC (proposer), ericp, Rob
16:08:30 [rob]
use case is
16:08:30 [DaveB]
this is proposing candidate requirements for the list, not supporting them
16:08:53 [DaveB]
kendall - complex graphs
16:09:16 [DaveB]
.. to select more than just based on an adjacent node such as three nodes connected by 2 edges, want just one?
16:09:55 [DaveB]
DanC - Pat Hayes said earlier, more than just paths?
16:10:06 [DaveB]
KC - paths with more than 1 edge
16:10:55 [DanC]
Jean Francois speaking, I believe
16:10:58 [DaveB]
16:11:04 [kendall]
taht's not me, btw
16:11:05 [DaveB]
JF - entailment as a requirement
16:11:31 [DaveB]
PH - QL should enable the querier to specify if they want interference or not
16:11:36 [DaveB]
16:12:35 [DaveB]
decided canidate requirement - paths of length more than two edges
16:12:59 [DaveB]
s/more than two/more than one/
16:13:25 [DaveB]
PatrickS - support PH inference
16:13:49 [DaveB]
AndyS - keep inference separate from source selection
16:14:25 [DaveB]
DanC - charter has words on this re inference, being same as selecting sources
16:14:58 [AndyS]
Inference Charter 1.8
16:15:13 [ericP]
16:17:13 [DaveB]
candidate requirement - should be possible for the query to indicate whether the response is expected to allow for entailment from the graph or is dealing with the graph as a fixed object (P Hayes)
16:17:34 [DaveB]
decided to add above, PatrickS agrees
16:17:42 [DaveB]
DanC - non trees?
16:17:58 [DaveB]
PH - more of an observation
16:18:30 [DaveB]
be able to ask if a triple doesn't exist in a grpah
16:19:01 [DaveB]
discussion of negation examples
16:20:35 [DaveB]
DanC - taking to email, subtlies here
16:21:07 [kendall]
"subtleties" is subtle ;>
16:21:31 [DaveB]
PatrickS - propose requirement to be able to express arbitrary datatype values
16:22:07 [thompsonbry]
16:22:40 [DaveB]
extend to 90 mins
16:23:07 [alberto]
60 - sorry need to leave
16:23:52 [DaveB]
? - my next Q would be operations on literal values, and then working on arbitrayr datatype values
16:24:07 [jannes]
(that was my)
16:24:24 [DaveB]
PatrickS - clearly there will be some limit tothe supported dataypes, such as XSD ones being widespread
16:24:54 [DaveB]
... for datatypes not supported, it could fail or could send back an error, don't understand
16:25:17 [DaveB]
patH - agree with patrickS
16:25:34 [AndyS]
AndyS agrees to the requirement
16:25:42 [DaveB]
janne - I agree also, min datatypes a QL should support?
16:25:45 [AndyS]
Issues on the actions on not supported
16:25:51 [DaveB]
?KC - agree, without the word arbitrary
16:25:56 [kendall]
nope, not me.
16:26:27 [DaveB]
candidate requirement above of PatrickS - decided on the list
16:26:28 [alberto]
AlbertoR agrees that core set of datatype support as relevant
16:26:59 [DaveB]
DanC - ask KC / EricP to copy these into the UC doc?
16:27:27 [Zakim]
- +1.207.395.aaaa
16:27:34 [ericP]
ACTION: EricP and KC to update use cases docuent with candidate requirement
16:28:21 [rob]
PS- queries should be bookmarkable
16:28:23 [thompsonbry]
Possible requirement: Client can negotiate with the server to determine which requirements MUST apply or MUST NOT apply for a given query.
16:28:44 [rob]
query captured as URL
16:28:53 [thompsonbry]
16:28:54 [rob]
alberto: concur, possible with XPointer
16:29:04 [rob]
new req: bookmarking
16:29:14 [thompsonbry]
Is that captured as the reference of a URI or as the URI itself?
16:30:11 [thompsonbry]
Suggestion is content negotiation for the returned representation?
16:30:14 [rob]
rob: users should be able to format their own results
16:30:26 [rob]
Eric- concur
16:31:23 [ericP]
candidate requirements += format their own results (specify format of return)
16:32:03 [rob]
Yanni: closure is useful
16:32:17 [DanC]
16:32:45 [kendall]
16:32:46 [rob]
Patrick supports
16:33:44 [thompsonbry]
Bryan supports client option to elect for closure, but also to elect for other content types in the response.
16:33:51 [rob]
Janne: closure implies ability to format results.
16:33:57 [rob]
(just an excuse to spell it correctly)
16:34:34 [rob]
the results can be an rdf graph...
16:35:09 [rob]
EricP: closure only if it means results as RDF
16:35:12 [DanC]
candidates += closure (Janne, EricP)
16:35:15 [DaveB]
I support results as rdf graph
16:35:44 [DaveB]
can somebody expand closure for me, as scribe, I've got to explain this!
16:36:14 [Zakim]
16:36:21 [alberto]
alberto has left #dawg
16:36:36 [jannes]
closure = operator that operates on data model B returns result in data model B
16:36:43 [rob]
proposed: ability to express queries in RDF
16:36:54 [DaveB]
thanks Janne
16:37:12 [DanC]
" ability to express queries in RDF" - candidate. jos, Patricks
16:37:18 [rob]
RDF vocab for expressing queries...
16:37:37 [kendall]
danc bottleneck! :>
16:38:01 [DanC]
Zakim, next item
16:38:01 [Zakim]
agendum 4. "Issues" taken up [from DanC]
16:38:33 [rob]
dan: there's mail not about use cases or requirements
16:38:41 [rob]
ignore, or put them in issues list?
16:38:50 [DanC]
16:39:06 [rob]
call for volunteers to manage issues list
16:39:13 [rob]
ericP volunteers
16:39:31 [rob]
eric and dan to maintain issues list
16:39:59 [DanC]
ACTION EricP: work with DanC on an initial issues list
16:40:53 [rob]
path: anything on issues list means lots of work; don't use it frivolously
16:41:02 [DanC]
Zakim, next item
16:41:02 [Zakim]
agendum 5. "Amsterdam meeting arrangements" taken up [from DanC]
16:41:09 [DanC]
16:41:37 [rob]
issues list should be approved by others?
16:41:42 [rob]
(who said it?)
16:41:44 [kendall]
so a candidates issues list :>
16:41:52 [kendall]
rob: patricks
16:41:57 [thompsonbry]
Signing off.
16:42:06 [Zakim]
16:42:07 [thompsonbry]
thompsonbry has left #dawg
16:42:34 [rob]
patricks: telecon for f2f include requirements
16:42:40 [rob]
(at least ps's requirements)
16:43:00 [DaveB]
can you expand on that?
16:43:13 [rob]
fill out reg form for f2f even if you're not attending
16:43:17 [rob]
all 17
16:43:58 [rob]
DanC to continue working on agenda for f2f
16:45:21 [Zakim]
16:45:24 [Zakim]
16:45:26 [rob]
fare thee well
16:45:29 [Zakim]
16:45:30 [DaveB]
16:45:31 [Zakim]
16:45:32 [Zakim]
16:45:39 [Zakim]
16:45:42 [Zakim]
16:45:50 [Zakim]
- +358.925.6aabb
16:45:54 [jannes]
jannes has left #dawg
16:45:54 [KevinW]
KevinW has left #dawg
16:47:30 [rob]
addition: PatS wanted the teleconference to be scheduled at the time the discussion of his requirements comes up.
16:47:41 [rob]
(during the face-to-face)
16:47:57 [rob]
Who's writing up minutes? Me, DanC, or DaveB?
16:48:01 [DaveB]
16:48:11 [DanC]
ah, good, dave, yes please.
16:48:18 [rob]
16:48:27 [rob]
Signing off, unless I'm needed for anything else...
16:48:35 [DaveB]
see ya
16:48:37 [DanC]
thanks, rob. hasta.
16:48:44 [Zakim]
16:52:50 [Zakim]
16:57:48 [DanC]
crud... did I ever get back to Brian?
17:02:03 [Zakim]
17:02:06 [Zakim]
17:02:07 [Zakim]
SW_DAWG()10:30AM has ended
17:02:08 [Zakim]
Attendees were Tayeb, [UMD], +1.207.395.aaaa, RobS, Kevin, +358.925.6aabb, DanielK, DanC, Pat_Hayes, PatrickS, +39.033.2.aacc, EricP, AlbertoG, [ASemantics], AndyS, BryanT, JosD
17:04:02 [DanC]
hmm... ericp, I wonder if the telcon reservation is for 15:30 UTC
17:04:23 [DanC]
"Boston *Thu 11:30 AM" --
17:05:17 [DanC]
crud... " 10:30am-12:00pm (14:30-16:00Z) SW_DAWG" --
17:05:46 [ericP]
kendal, issues list:
17:08:51 [DanC]
ericp, pls put that under 2001/sw/DataAccess
17:09:45 [ericP]
ok. kendal, i'm moving that document now
17:12:43 [ericP]
kendal, now at
17:12:48 [DanC]
17:20:20 [kendall]
kendall has joined #dawg
17:58:20 [patH]
18:12:46 [kendall]
ericp: you still here?
18:15:19 [ericP]
18:15:23 [kendall]
oh, hey
18:15:30 [kendall]
hmm, i'm a bit confused by the uri you coined
18:15:35 [kendall]
... /issues?
18:15:52 [ericP]
oops, too many things in my head at once.
18:15:54 [kendall]
that sounds like the issues list. but the doc i'm working on isn't an issues list. :>
18:16:00 [ericP]
i needed to create an issues list too.
18:16:07 [kendall]
maybe .../user-stories or something?
18:16:08 [kendall]
18:16:10 [kendall]
i thought so
18:16:24 [ericP]
UseCases.html ?
18:16:34 [kendall]
well, no .html but sure ;>
18:17:48 [DanC]
RRSAgent, make logs world-readable
18:17:51 [DanC]
RRSAgent, make logs world-access
18:18:08 [DanC]
RRSAgent, pointer?
18:18:08 [RRSAgent]
18:18:45 [ericP]
18:19:29 [DanC]
are you fixing kendall up with cvs access, ericp?
18:21:19 [ericP] is now an empty issues list
18:21:45 [ericP]
danc, kendall and i decided to start with him using mail to me as his publishing mechanism
18:22:03 [ericP]
and deal with more access if necissary
18:22:21 [kendall]
well, mail to the list ;>
18:22:27 [ericP]
specifically, mail to the list, which i will poll and publish in the consistent location
18:22:31 [ericP]
18:22:37 [kendall]
18:31:54 [DanC]
18:32:13 [DanC]
I'd really rather kendall had instant-update-access.
18:32:25 [DanC]
we can move the document to mindswap if it helps.
18:33:39 [kendall]
i dont' really have much of a preference
18:34:06 [DanC]
ok, well, I have a pretty strong preference that you use directly, kendall
18:34:33 [kendall]
18:38:59 [ericP]
danc, may i ask why? are you worried about letancy or burdens falling on your shoulders when i'm off net?
18:39:38 [DanC]
(a) I don't like whole documents being mailed to the whole WG to fix a typo. And I want typos fixed quickly and often.
18:39:47 [ericP]
we didn't think it was worth a lot of infrastructure effort at this point, but you may have a better grasp of the job than either of us.
18:39:47 [DanC]
(b) I don't want 2 people in the critical path when 1 will do.
18:40:15 [DanC]
I want WG members to get instant-gratification for proposing changes to the document.
18:40:32 [ericP]
all right, i'll start the dance with sysreq tomorrow
18:40:54 [DanC]
as in: "ah, yes, good point. fixed in $Revision 1.6$ of"
18:41:44 [DanC]
thanks, ericp
18:41:53 [ericP]
no prob
18:51:10 [Zakim]
Zakim has left #dawg
19:02:25 [DanC]
RRSAgent, bye
19:02:25 [RRSAgent]
I see 3 open action items:
19:02:25 [RRSAgent]
ACTION: EricP to Send mail to group about rdf-query-rules doc, proposing to using it an abstract syntax to describe requirements [1]
19:02:25 [RRSAgent]
recorded in
19:02:25 [RRSAgent]
ACTION: EricP and KC to update use cases docuent with candidate requirement [2]
19:02:25 [RRSAgent]
recorded in
19:02:25 [RRSAgent]
ACTION: EricP to work with DanC on an initial issues list [3]
19:02:25 [RRSAgent]
recorded in