IRC log of rdfcore on 2002-07-26

Timestamps are in UTC.

13:48:22 [RRSAgent]
RRSAgent has joined #rdfcore
13:48:32 [em]
agenda +Volunteer scribe
13:48:41 [em]
agenda +Roll Call
13:48:52 [em]
Agenda +Review Agenda - http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0155.html
13:49:04 [em]
agenda + Next telecon Aug 2nd 2002
13:49:28 [em]
agenda +Review minutes of 2002-07-19 teleconference - (no minutes but log http://www.w3.org/2002/07/19-rdfcore-irc)
13:49:42 [em]
agenda +action item status
13:49:55 [danbri]
danbri has joined #rdfcore
13:50:00 [em]
agenda +rdf:ID / rdf:node proposal - http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0080.html
13:50:14 [ilrt]
ilrt has joined #rdfcore
13:50:18 [em]
agenda +New Document - http://www.ninebynine.org/wip/RDF-basics/2002-07-25/Overview.html
13:50:31 [em]
agenda +rdfns-assertion
13:50:42 [em]
agenda +Procedure for determining reserved vocabulary
13:50:57 [em]
agenda +Schedule and Process to Last Call
13:51:01 [em]
agenda +datatypes
13:51:21 [ilrt2]
ilrt2 has joined #rdfcore
13:56:08 [gk-scribe]
gk-scribe has joined #rdfcore
13:57:43 [em]
zakim, list conferences.
13:57:44 [Zakim]
I see SW_RDFCore()10:00AM, WAI_EOWG()8:30AM
13:57:53 [em]
zakim, this is SW_RDFCore.
13:57:54 [Zakim]
sorry, em, I do not see a conference named 'SW_RDFCore.'
13:57:57 [em]
zakim, this is SW
13:57:58 [Zakim]
ok, em
13:58:05 [em]
zakim, who is here?
13:58:06 [Zakim]
On the phone I see EricM
13:58:07 [Zakim]
On IRC I see gk, ilrt2, ilrt, danbri, RRSAgent, Zakim, em, logger_1
14:00:45 [Zakim]
+??P3
14:01:01 [Zakim]
+Manola
14:01:02 [Zakim]
+??P11
14:02:57 [em]
zakim, who is here?
14:02:58 [Zakim]
On the phone I see EricM, ??P3, Manola, ??P11
14:02:59 [Zakim]
On IRC I see gk, jang, DaveB, danbri, RRSAgent, Zakim, em, logger_1
14:03:25 [jang]
zakim, don't pick me!
14:03:26 [Zakim]
I don't understand 'don't pick me!', jang. Try /msg Zakim help
14:03:56 [danbri]
zakim, exempt.add(danbri,'wrists hurt')
14:03:57 [Zakim]
I don't understand 'exempt.add(danbri,'wrists hurt')', danbri. Try /msg Zakim help
14:04:26 [Zakim]
+??P13
14:04:32 [Zakim]
+DanBri
14:04:40 [em]
zakim, +??P13 is jjc
14:04:41 [Zakim]
sorry, em, I do not recognize a party named '+??P13'
14:05:02 [jang]
zakim, ??P13 is jjc
14:05:04 [Zakim]
+Jjc; got it
14:05:06 [Zakim]
+??P12
14:05:08 [connolly]
connolly has joined #rdfcore
14:05:37 [em]
agenda?
14:06:11 [DaveB-scr]
jjc scribe
14:06:14 [em]
zakim, who is here?
14:06:15 [Zakim]
On the phone I see EricM, ??P3, Manola, ??P11, DanBri, Jjc, ??P12
14:06:15 [DaveB-scr]
(daveb doing for now)
14:06:15 [Zakim]
On IRC I see connolly, gk, jang, DaveB-scr, danbri, RRSAgent, Zakim, em, logger_1
14:06:46 [em]
gk is here
14:06:53 [em]
stevep is here
14:06:58 [em]
daveb and jan are also here
14:07:02 [gk]
zakim, ??P11 is GK
14:07:04 [Zakim]
+GK; got it
14:07:05 [DaveB-scr]
rollcall complete
14:07:09 [DaveB-scr]
review agenda
14:07:25 [DaveB-scr]
item 4
14:07:43 [DaveB-scr]
donm, 2 august
14:07:50 [jjcscribe]
jjcscribe has joined #rdfcore
14:07:57 [jang]
reg bwm (holidaY0 jos \
14:08:08 [jjcscribe]
I am on now.
14:08:18 [jjcscribe]
reg danc
14:08:19 [DaveB-scr]
regrets josd
14:08:28 [jang]
[excuse poor trying, imac ketbroad]
14:08:39 [jjcscribe]
Minutes Last Telecon
14:08:53 [jjcscribe]
Patrick agreed to do minutes but they are missing
14:09:09 [jjcscribe]
ACTION ericm Chase minutes of last telecon
14:09:20 [jjcscribe]
Current record is IRC log
14:10:30 [Zakim]
+Mike_Dean
14:10:46 [jjcscribe]
ACTION danbri versus LBase
14:11:04 [jjcscribe]
dan will circulate something before Wednesday
14:11:37 [jjcscribe]
This will be a note coming out of the WG discussion
14:12:32 [jjcscribe]
ACTION: jang update test cases - continued
14:12:50 [jjcscribe]
ACTION: eric Look into why jang ... continued
14:13:47 [jjcscribe]
2002-06-28# closed
14:13:55 [jjcscribe]
Because close
14:14:56 [Zakim]
+??P15
14:15:17 [jjcscribe]
Jos joined
14:15:20 [jjcscribe]
Agenda item 7
14:15:23 [DaveB]
regrets were from PatH
14:16:01 [jjcscribe]
New agenda item 7a rdf:ID
14:16:08 [jang]
jjc:
14:16:17 [jang]
my take is that xml NS production is the correct one
14:16:24 [jang]
ie, a ns name, without the colon
14:16:34 [jang]
djb: we've been over this, can't we decide it?
14:16:36 [danbri]
Oh., em, Agenda request: Namespaces 1.1 LC review.
14:16:38 [jang]
dave: anyone object?
14:16:49 [jang]
frankm: what's the issue again?
14:16:50 [em]
ok, danbri so noted
14:16:57 [gk]
NCName
14:16:58 [gk]
::=
14:16:58 [gk]
(Letter | '_') (NCNameChar)*
14:16:58 [gk]
An XML Name, minus the ":" */
14:16:59 [em]
agenda +Namespace 1.1
14:17:00 [jang]
jjc: test case sent to wg the other day... just looking for it
14:17:30 [jang]
daveb: the issue is that xml id and rdf id take different sets of characters
14:17:33 [jang]
they shold be the same
14:17:39 [jang]
then xml schema can validate it
14:17:45 [jang]
em: objections?
14:17:58 [danbri]
re aggendum 13, see Namespaces 1.1 LC plan: http://lists.w3.org/Archives/Member/chairs/2002JulSep/0040.html (need W3C Member passwd)
14:18:07 [jang]
decision: this is obvious, we should do this?
14:18:33 [jang]
jjc points to message 0148 from this month
14:19:24 [DaveB]
xml id and rdf id in the syntax take differetn set of charcater
14:19:30 [DaveB]
I propose that we make them the same
14:19:39 [DaveB]
and then we can use w3x xsl to validate it.
14:20:00 [jang]
mike: issues with some tools: ids can't begin with numbers?
14:20:06 [jang]
currently allowed by rdf, not xml
14:20:11 [jang]
daveb: that's the change
14:20:28 [jang]
miked: rdf is becoming more restrictive here, right?
14:20:32 [jjcscribe]
mike: rdf is becoming more restrictive
14:20:53 [jjcscribe]
consider ssn or zip code
14:21:16 [em]
q+
14:22:22 [DaveB]
connolly: was ambiguous, as jjc explaining on the telcon
14:22:28 [DaveB]
can't find ptrs to more just this sec
14:22:50 [jjcscribe]
Abstentions: miked
14:23:00 [jjcscribe]
Proposal carried.
14:23:10 [jjcscribe]
ACTION jjc update test case
14:23:24 [DaveB]
(jjc: can you split test cases so one thing tested rather than two?)
14:23:41 [jjcscribe]
ACTION dave update syntax
14:23:53 [mdean]
mdean has joined #rdfcore
14:23:57 [JosD]
JosD has joined #rdfcore
14:24:14 [jjcscribe]
ACTION jjc include test case with numeric ID
14:24:20 [DaveB]
connolly: the three lines I typed higher up "xml id ..." up to "validate it"
14:24:54 [jang]
item 7b
14:25:03 [jang]
jjc has simple proposal for rdf:node
14:25:09 [jang]
doesn't address all aspect sof the syntax
14:25:21 [jang]
but it solves the bulk of the roundtripping problem
14:25:28 [jang]
a number of choicepoints highlighted in the email...
14:25:35 [jang]
..we could discuss those here... chair?
14:25:46 [jang]
em: want to get this finished.
14:25:56 [jang]
that being said, don't know if people consider this a no-brainer
14:26:09 [jang]
one choice-point seems obvious, the rest less so
14:26:28 [jang]
chair's position was not to reopen at this time, but if a slam-0dunk then reopen and finish it
14:26:32 [danbri]
this= rdf:node idea?
14:26:34 [jang]
is this a slam dunk?
14:26:38 [DaveB]
yes
14:26:39 [jang]
yes
14:26:45 [jang]
(yes to danbri)
14:26:48 [DaveB]
(not agreeing, replying to danbri)
14:27:02 [jang]
jjc: there is a proposal
14:27:09 [jang]
which is moderately long but is complete, I think
14:27:23 [jang]
that we add the file-scope identifier for a blank node as an attribute
14:27:34 [jang]
we can label subject or object with such an attribute.
14:27:39 [jang]
that's about it
14:27:44 [em]
danbri, please do not deal with the comments thing - please focus on the meeting at hand
14:27:44 [danbri]
(rdf:node is what rdf:ID should've been)
14:27:49 [jang]
em: comments?
14:28:03 [jang]
frank: question: concerning the use of xml strings as blank-node identifiers?
14:28:16 [jang]
jjc: i could have been more restrictive tthan just @xml strings@
14:28:32 [jang]
daveb; coincidentally, the ntriples IDs are also xml ids
14:28:46 [jang]
frank: how do you distinguish it as a blank node id?
14:28:53 [jang]
danbri: a new attribute for it
14:29:15 [jang]
daveb: shouldn't look like a uri
14:29:25 [jang]
should look like rdf:id
14:29:33 [jang]
jjc: ntriples is us-ascii,
14:29:43 [jang]
we should allow id's as e-acute, etc.
14:29:55 [jang]
danbri: ntriples is just a test format
14:30:10 [jang]
jjc: there's no reason why these have _anything_ to do with ntriples identifiers
14:30:18 [jang]
it's just another syntax
14:30:30 [jang]
if the id's don't survive roundtripping because they just label a blank node...
14:30:36 [jang]
not a problem
14:30:44 [jang]
em: what's the risk if we don't have this?
14:30:57 [jang]
jjc: there are real users who really want to write rdf/xml that they've read in
14:31:04 [jang]
in a programmatic sense
14:31:16 [jang]
and at some point they need blank node ids to do that in certain circumstances
14:31:24 [jang]
if they assign a uri then the meaning has changed
14:31:31 [jang]
they might do something nonstandard
14:31:44 [jang]
in jena, we're feeling increasing pressure to do something nonstandard
14:31:50 [jang]
which we'd rather not do
14:32:10 [jang]
danbri: ntriples are increasingly attractive because they're the only roundtrippable format
14:32:23 [jang]
jjc: if we don't do this then the user community is being effectively forced to use ntriples
14:32:29 [jang]
even though we don't endorse it
14:32:30 [danbri]
zakim, q+ rdf:node
14:32:31 [Zakim]
I see Em, Rdf:node on the speaker queue
14:32:34 [danbri]
doh
14:32:39 [em]
ack em
14:32:40 [danbri]
zakim, q- rdf:node
14:32:41 [Zakim]
I see no one on the speaker queue
14:32:42 [jang]
\frank: the roundtripping is rdf/xml -> somehitng, -> rdf/xml
14:32:49 [danbri]
zakim, q+
14:32:50 [Zakim]
I see Danbri on the speaker queue
14:33:05 [jang]
I want to read rdfxml, store it, generate it
14:33:13 [jang]
if blank node ids are allowed to look like uris?
14:33:29 [jang]
what i have to do (i think) is adopt some nonstandard approach to store these?
14:33:38 [jang]
danbri: it's inside your db implementation
14:33:42 [gk]
If we adopt them, I think graph-scope IDs should be limited to N-triple allowed form, I18N not needed because its a machine notion, not human-readable. This seems simplest effective approach.
14:33:50 [jang]
jjc: jena has effectively a bit on every resource indicating the label type
14:34:40 [jang]
danbri: have same syntactic constraints as we do on rdf:id
14:34:50 [jang]
there are few things I've really cared about as an implementor
14:34:54 [jang]
this is one of those
14:34:57 [gk]
I think the compelling case for this feature is to provide applications a standard way to communicate arbitrary graphs.
14:34:59 [jang]
I can load, query, merge, etc. rdf
14:35:03 [jang]
but I can't dump it back as a file
14:35:09 [jang]
it's honestly embarrassing
14:35:35 [jang]
em: unless we standardise ntriples, with i18n etc. this kind of mechanism is necessary
14:35:38 [DaveB]
maybe rdf:nodeID ?
14:35:42 [jang]
(yesses )
14:35:45 [danbri]
yes, I like rdf:nodeID
14:35:56 [jang]
em: I like danbri's suggestion of synchronising with rdf:id
14:36:01 [danbri]
This is really the continuation of our cleanup re anonymous/blank resources.
14:36:07 [jang]
jjc: I'm happy with node id
14:36:17 [gk]
q+
14:36:26 [jang]
then we restrict to string syntax of rdf:id (= xml:id)
14:36:28 [jjcscribe]
rdf:nodeID
14:36:39 [jang]
daveb: frank - much rewriting on the primer?
14:36:45 [jang]
frank: er, there will be some...
14:36:57 [jang]
I'd like a clear explanation of what the change is...
14:37:05 [jang]
in particular, what these node ids are going to look like
14:37:14 [jang]
em: premature, i think ,without that being written up
14:37:24 [connolly]
er... nodeID and nodeRef? or do you use nodeID in both places? note you can't use rdf:resource to refer to these things.
14:37:28 [jang]
davbe: isn't that's jjc's proposal plus the amendments we're discussing?
14:37:33 [DaveB]
connolly: rdf:nodeID in both
14:37:36 [connolly]
ew
14:37:48 [em]
-> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0080.html
14:37:49 [connolly]
oh well
14:37:56 [DaveB]
haven't decided yet DanC
14:38:25 [jang]
(quick rummage through mailing lists...)
14:38:39 [jang]
frankm: this doesn't mention what's in the primer about blank nodes
14:38:56 [jang]
because we're talking about generating those ids, not explicitly identifying them
14:39:11 [jang]
jjc: the proposal is that the blank nodes remain blank
14:39:24 [jang]
we just want to allow a file to refer to the same blank node in two places\
14:39:40 [jang]
there's some push to allow blank nodes with a @global@ scope identifier. that's _not_ the proposal
14:39:55 [jjcscribe]
the playground is me ...
14:40:00 [jang]
frankm: you're being allowed to explicitly specify an id for a blank node
14:40:02 [jang]
right?
14:40:13 [jang]
secondly: are you always required to specify an id for a blank node?
14:40:24 [jang]
or are the current syntactic abbrevs still allowed?
14:40:34 [danbri]
hmm, rdf:nodeID seems ok for having something to point to. But how (re dan's questoin) do we point to it?
14:40:51 [em]
q?
14:40:58 [danbri]
q-
14:40:59 [DaveB]
I don't mind, rdf:resourceID/resourceRef - hmm
14:41:15 [em]
ack gk
14:41:18 [jang]
jjc: second question: the answer is not required
14:41:31 [jang]
jang: first question: no suggestion that blank node id -strings_ survive a roundtripping
14:41:43 [danbri]
The proposal doesn't mention resourceID or similar; is the proposal incomplete?
14:41:56 [jang]
gk: jjc's said it's not necessary to have different names for rdf:nodeid and rdf:nodeidref
14:42:06 [jang]
but is it worth having two attributes for subject and object?
14:42:21 [jang]
jjc: the proposal is to use the same attribute twice
14:42:49 [jang]
danbri; I've often wished rdf:about and rdf:resouce were just rdf:webid
14:42:57 [jang]
jjc: it makes the striping easier to see
14:43:08 [jang]
taking the opposite viewpoint is maybe perverse...?
14:43:12 [danbri]
I like it.
14:43:19 [jang]
gk; both work technically, just from a pedagogical pov, which is better?
14:43:28 [jang]
nodeid and nodeidref work for me (gk)
14:43:31 [DaveB]
gk mentions rdf:nodeIDref
14:43:32 [DaveB]
oh
14:43:43 [jang]
jjc: my muse deserted me, i only came up with one name
14:44:35 [jang]
(the "bnode" terminology is almost resurrected... but not)
14:44:52 [jang]
em: is the view of the group this is (a) important, and (b) this is the way to do it?
14:45:01 [jang]
decisions in then to get this into the spec, right?
14:45:16 [jang]
em: this puts timescales at risk, so if we can agree...
14:45:20 [jang]
and I think we have...
14:45:25 [jang]
let's make the decision now, ok?
14:45:35 [jang]
otherwise, this is no longer a slam-dunk
14:45:41 [jang]
(frank: it's a three-point-shot)
14:45:55 [jang]
jjc: my full resolution of msg 0080
14:45:59 [jang]
with node replaced by nodeid
14:46:10 [jang]
and nodeid attr value the same restrictions as rdf:id attr value?
14:46:37 [jang]
that should be rdf:nodeID
14:46:40 [gk]
Do we have separate attribute for subj/obj?
14:46:43 [DaveB]
no
14:47:00 [jang]
jjc: personally, I prefer not
14:47:04 [jang]
danbri: not
14:47:37 [jang]
mike: symmetry argument?
14:47:42 [jang]
with resource, about?
14:48:20 [jang]
jjc asks for suggestions.. nodeabout, noderef, bid, babout, etc.
14:48:28 [jang]
none meet with much agreement
14:48:39 [jang]
s/much/any
14:49:08 [jang]
em: stick with one. it'll hurt, but big payoff
14:49:14 [jang]
with two, hurts much more for the same benefit
14:49:51 [em]
zakim, who is here?
14:49:52 [Zakim]
On the phone I see EricM, ??P3, Manola, GK, DanBri, Jjc, ??P12, Mike_Dean, ??P15
14:49:53 [Zakim]
On IRC I see JosD, mdean, jjcscribe, DanC, gk, jang, DaveB, danbri, RRSAgent, Zakim, em, logger_1
14:50:03 [jang]
yesses to the revised proposal above...
14:50:09 [jang]
em asks for no's?
14:50:12 [jang]
abstentions?
14:50:19 [jang]
none and none, respectively
14:50:34 [jang]
agreement. congrats, jeremy.
14:50:45 [jang]
(the actions are in the resolution...)
14:50:57 [jang]
action: daveb to update syntax doc to reflect nodeID
14:51:04 [jang]
action: jjc to produce test cases
14:51:08 [jang]
(on nodeID)
14:51:21 [jang]
action daveB to update syntax to bring rdfid into line with xmlid
14:51:31 [gk]
See: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0159.html
14:51:37 [jang]
[want to scribe again jjc?]
14:51:52 [jjcscribe]
action daveB to update rdfms-names-use to reflect rdf:nodeID
14:53:22 [jjcscribe]
Graham talks through new doc.
14:53:36 [jjcscribe]
em: the section 2.3 needs activ review
14:53:52 [jjcscribe]
(next agenda item)
14:54:29 [jjcscribe]
Where Dave has suggested pointing to syntax doc we will,
14:54:40 [jjcscribe]
but unclear exactly where
14:54:48 [jang]
[note: could do with this living somewhere under w3-space so that TC doc can refer to it normatively... in the near future]
14:55:07 [jjcscribe]
Section 4.2 fragment IDs is new material that needs atcive WG agreement
14:56:00 [danbri]
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Jul/0159.html points to a 404
14:56:04 [jang]
[unavailable for review, out of the country next week, sorry]
14:56:47 [DaveB]
I asked gk to post to www-archive, should be a copy there
14:57:05 [jjcscribe]
Isn't that done?
14:57:24 [jjcscribe]
em: is this ready to go out as a WD?
14:57:31 [jjcscribe]
gk: yes, jjc: nearly
14:57:39 [DaveB]
gk's doc - http://lists.w3.org/Archives/Public/www-archive/2002Jul/att-0052/01-Overview.htm
14:57:44 [DaveB]
gk/jjc's 25/ july doc
14:57:50 [danbri]
(beat me to it, ta)
14:58:15 [DaveB]
q+
14:58:22 [jjcscribe]
ACTION gk release doc on Tuesday
14:58:26 [DaveB]
hmm, no Zakim - nevermind
14:58:52 [jjcscribe]
who can review this for Friday?
14:58:52 [Zakim]
Zakim has joined #rdfcore
14:59:07 [danbri]
I re-invited Zakim; don't know why it left, or whether it lost state.
14:59:10 [danbri]
zakim, who is here?
14:59:11 [Zakim]
sorry, danbri, I don't know what conference this is
14:59:12 [Zakim]
On IRC I see Zakim, JosD, mdean, jjcscribe, DanC, gk, jang, DaveB, danbri, RRSAgent, em, logger_1
14:59:24 [gk]
BTW, proposed title: Miscellanea and Abstract Data Model
14:59:38 [jang]
RDF MADM?
15:00:09 [em]
q?
15:00:25 [jjcscribe]
DaveB has pretty much reviewed it.
15:00:39 [jjcscribe]
Are there two more?
15:00:47 [jang]
[needs rdf:nodeID in section 3.7]
15:00:48 [jjcscribe]
ACTION ericm Review document.
15:01:14 [jjcscribe]
3.7 could be in syntax doc?
15:01:24 [jjcscribe]
ACTION danbri Review document.
15:01:33 [jang]
yep
15:01:50 [danbri]
(noting that I'm not 100% confident I'll manage it, but want to and will try my best to...)
15:02:02 [DaveB]
em: timecheck
15:03:02 [jjcscribe]
ACTION jos Partial review (emphasis on section 2.3)
15:03:13 [jjcscribe]
ACTION danbri Review of section 2.3
15:03:26 [danbri]
ie. http://lists.w3.org/Archives/Public/www-archive/2002Jul/att-0052/01-Overview.htm#section-Meaning
15:03:53 [jjcscribe]
ACTION ericm Solicit reviews on rdf-core wg
15:04:03 [jjcscribe]
ACTION frank Review section 2.3
15:05:35 [jjcscribe]
em: last call scheculde - problem is datatypes.
15:07:20 [jang]
bye folks, thanks.
15:07:28 [jjcscribe]
Possible theme for next week is datatypes.
15:07:35 [jjcscribe]
adjourned.
15:07:59 [gk]
DaveB, can you point me to vocab list in syntax doc pls?
15:08:10 [DaveB]
hold on
15:09:36 [gk]
Ta .. (I lose track of where to find the latest working version)
15:10:12 [gk]
zakim, who is her5e?
15:10:13 [Zakim]
I don't understand your question, gk.
15:10:16 [gk]
zakim, who is here?
15:10:17 [Zakim]
sorry, gk, I don't know what conference this is
15:10:18 [Zakim]
On IRC I see Zakim, JosD, jjcscribe, DanC, gk, DaveB, danbri, RRSAgent, em, logger_1
15:10:43 [danbri]
zakim loststate when dropped off channel.
15:10:52 [danbri]
anyone know what happend?
15:11:54 [DaveB]
3.4 http://www.w3.org/TR/rdf-syntax-grammar/#section-Namespace
15:11:58 [DaveB]
is where rdf namespace defined
15:12:00 [DaveB]
and has all the terms
15:12:10 [DaveB]
in the editor's draft, there are expansions and updates
15:13:58 [DaveB]
editor's draft version http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/#section-Namespace
15:14:10 [DaveB]
with collection stuff - first, rest, nil
15:15:36 [DaveB]
end of chat
15:23:06 [DanC]
DanC has left #rdfcore
17:38:00 [Zakim]
Zakim has left #rdfcore
17:38:48 [em]
em has left #rdfcore
17:46:20 [danbri]
danbri has left #rdfcore