IRC log of rdfcore on 2003-06-20

Timestamps are in UTC.

14:02:24 [RRSAgent]
RRSAgent has joined #rdfcore
14:02:31 [Zakim]
+DanC
14:02:49 [DanC]
agenda?
14:03:51 [bwm]
agenda: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0134.html
14:04:11 [DanC]
agenda + 20Jun http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0134.html
14:04:25 [Zakim]
+??P18
14:05:37 [DanC]
DanC has changed the topic to: rdfcore jun 20 telecon. scribe: jjc
14:07:35 [jjcscribe]
agendum 1: scribe is jjc
14:08:07 [jjcscribe]
Zakim, who's on the phone?
14:08:07 [Zakim]
On the phone I see jjc, FrankM, bwm, DanC, ILRT
14:08:08 [Zakim]
ILRT has JanG, DanBri
14:08:09 [bwm]
zakim, who is on the phone?
14:08:09 [Zakim]
On the phone I see jjc, FrankM, bwm, DanC, ILRT
14:08:10 [Zakim]
ILRT has JanG, DanBri
14:08:21 [DanC]
agenda + warm-up calesthetics or do the wave or something
14:08:38 [bwm]
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0151.html
14:08:53 [jjcscribe]
agendum item 2 roll call complete
14:09:06 [jjcscribe]
agendum item 3 review agenda
14:09:23 [jjcscribe]
the chair plans to tkae AOB before item 15
14:09:27 [Zakim]
Not knowing who is chairing or who scribed recently, I propose DanC
14:09:30 [jjcscribe]
4 next meeting
14:09:33 [Zakim]
Not knowing who is chairing or who scribed recently, I propose JanG
14:09:36 [jjcscribe]
no scribe volunteer
14:09:39 [Zakim]
Not knowing who is chairing or who scribed recently, I propose FrankM
14:10:05 [Zakim]
+EMiller
14:10:33 [jjcscribe]
DanC volunteers to scribe, EM will cover
14:10:48 [jjcscribe]
agendum 5 minutes approved
14:10:51 [jjcscribe]
agendum 6 minutes approved
14:11:20 [DanC]
oops; neglected to check "update concepts to reflect disposition of danc-01"
14:11:22 [jjcscribe]
agendum 7 complete actions recorded
14:11:27 [DanC]
I guess I'll get a formal "are you ok?"
14:12:28 [jjcscribe]
agendum 8 witrhdrawn actions
14:12:42 [DanC]
(is there a test that shows synonyms for XMLLiteral? maybe it can't be done)
14:13:38 [jjcscribe]
jjc: note that synonyms for XMLLiterals are allowed
14:13:46 [jjcscribe]
http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0173.html
14:13:54 [jjcscribe]
agendum 9 OWL test cases
14:14:00 [jjcscribe]
see msg above,
14:14:07 [jjcscribe]
ACTION Jan review OWL Test Cases
14:14:17 [DanC]
jjc's msg to webont http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0173.html
14:14:18 [jjcscribe]
ACTION jjc clairfy scope of review of OWL Test Cases
14:14:56 [jjcscribe]
agendum 10 value space of XMLLiteral
14:16:49 [gk]
gk has joined #rdfcore
14:17:06 [gk]
gk has joined #rdfcore
14:17:16 [gk]
Oops, sorry I'm late. Dialling soon.
14:17:27 [bwm]
welcome graham
14:18:49 [Zakim]
+GrahamKlyne
14:19:45 [jjcscribe]
jumping to agenda item 13 I18N comment
14:19:55 [jjcscribe]
(jjc has explained what exlcussive canoncial XML is)
14:21:04 [jjcscribe]
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0150.html
14:21:21 [bwm]
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0025.html
14:21:25 [jjcscribe]
(above msg contains exclusive canoncial XML discussion)
14:21:28 [DanC]
the agenda doesn't cite the most relevant msg from I18N, to me http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0023.html
14:21:37 [bwm]
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0132.html
14:23:59 [JosD]
JosD has joined #rdfcore
14:24:32 [Zakim]
+??P14
14:24:54 [JosD]
Zakim. ??P14 is JosD
14:25:37 [gk]
q+ to say that having to look inside the literal to look for markup is very painful to me
14:29:25 [em]
is daveb here?
14:29:46 [jjcscribe]
discussion of I18N msg
14:30:17 [jjcscribe]
should treatmnet of parseType="Literal" be dependent on markup in side string
14:31:53 [jjcscribe]
q+ to mention Cannes, and problems of not being able to distinguish
14:32:26 [jjcscribe]
dan suggests folliowng I18N suggestion ..
14:32:33 [jjcscribe]
^dan^danc
14:33:18 [jjcscribe]
bwm clarifies that dan is proposing a change to the parser
14:35:02 [jjcscribe]
bwm asks can we reject this on the graounds of lateness?
14:35:21 [jjcscribe]
danc maybe ...
14:36:20 [gk]
Is an XML literal *really* a *byte* sequence, unlike a string??
14:36:24 [danbri_dna]
q+ to have jang ask about something
14:37:06 [gk]
q-
14:37:56 [jjcscribe]
danc: the I18N requirement is clear
14:38:21 [bwm]
q?
14:38:27 [gk]
q+ to make two suggestions: (a) restrict parsetype=literal, (b) drop XMLLiteral
14:38:35 [jjcscribe]
danc: we need to talk to them some more if we choose not to address the req
14:38:58 [bwm]
ack danbri
14:38:58 [Zakim]
danbri_dna, you wanted to have jang ask about something
14:39:43 [jjcscribe]
jan asks why not use strings in value space?
14:39:56 [jjcscribe]
jjc I think there was a reason but I can't remember it
14:40:11 [bwm]
ack gk
14:40:11 [Zakim]
gk, you wanted to make two suggestions: (a) restrict parsetype=literal, (b) drop XMLLiteral
14:40:15 [bwm]
q?
14:40:19 [bwm]
ack jjscribe
14:40:34 [jjcscribe]
q+ to explain why strings are not good
14:42:14 [jjcscribe]
gk (a) could we restrict =literal to case when there is markup
14:43:20 [jjcscribe]
jang: difficulties in serialization
14:44:16 [bwm]
ack jjc
14:44:16 [Zakim]
jjcscribe, you wanted to mention Cannes, and problems of not being able to distinguish and to explain why strings are not good
14:45:17 [gk]
I think I disagree wioth jjc's entailment
14:45:30 [DanC]
I hope it's in the test suite
14:45:50 [gk]
DanC, I think it certainly should be
14:46:50 [gk]
*is* XML a tree structure, or does in *encode* a tree structure
14:47:29 [gk]
DanC says it's important to distinguish between a string that looks like XML, and "real" XML
14:48:18 [gk]
GK asks: is it also important to distinguish between a string that happens to contain digits, and a string that denotes a number?
14:48:20 [DanC]
that's the motivation for parseType="Literal", no?
14:48:47 [gk]
DanC, that's not what I understood Ralph was claimed to say
14:49:06 [DanC]
well, it's my motivation for parseType="Literal"
14:49:14 [jjcscribe]
jos: can we use the datatype
14:50:27 [jjcscribe]
does distinguishing XML and non-XML conflict with not distinguishing text from text?
14:50:36 [jjcscribe]
gk: yes. danc: no.
14:51:34 [gk]
q+ to offer another design
14:51:50 [DanC]
consider <prop parseType="Literal">foo</> vs <prop>foo<//>
14:51:57 [DanC]
^bwm's example
14:53:13 [DanC]
jjc: consider <prop parseType="Literal">foo &amp; bar </> vs <prop>foo &amp; bar</>
14:54:05 [gk]
q-
14:56:25 [jjcscribe]
fm: parseTytpe="Literal" implicitly makes datatype = "XMLLiteral"
14:57:25 [gk]
q+ to ask if parsetype=literal doesn't of itself constitute a different representation
14:59:02 [jjcscribe]
q+ to ask about charmod
15:00:17 [DanC]
jjc/fm: consider <prop parseType="Literal">100</> where upstream groks xsd:integer. [?]
15:01:55 [DanC]
how does this motivate different abstract syntax from <prop>100</>
15:02:03 [DanC]
?
15:02:20 [jjcscribe]
jjc: perhaps not very compelling
15:03:26 [gk]
Jos suggests if you *really* want xmlliteral, then specify datatype XMLLiteral (?)
15:03:40 [bwm]
jjc: consider <prop parseType="Literal">foo &amp; bar </> vs <prop>foo &amp; bar</>
15:04:01 [jjcscribe]
danc: the I18N group has not convinced RDF Core
15:04:05 [jjcscribe]
q+
15:04:56 [DanC]
... not convinces RDF Core to accept the requirement that <prop parseType="Literal">abc</> and <prop>abc</> have the same denotation
15:05:03 [DanC]
convinced
15:08:32 [gk]
i.e. <foo>abc</foo> and <foo rdf:parsetype="Literal">abc</foo> are different representations.
15:08:36 [DanC]
I'm not comfortable with "representation" here.
15:08:42 [em]
i'm here... unmuted to speak but will follow up via irc
15:09:03 [bwm]
ack gk
15:09:03 [Zakim]
gk, you wanted to ask if parsetype=literal doesn't of itself constitute a different representation
15:09:30 [jjcscribe]
fm: are plain literals same as xsd:string?
15:09:36 [jjcscribe]
some yeses, one agnostic
15:09:43 [em]
i had this on an agenda with tim yesterday to help prep for this meeting (in case this happened)... unfortunatly didnt get to this due to other agenda items. I agree wrt getting a sense on management here - willig to do this
15:10:10 [em]
q+ to confirm the previously raised issue about contradictory information
15:10:19 [bwm]
ack jjc
15:10:19 [Zakim]
jjcscribe, you wanted to ask about charmod and to
15:10:27 [bwm]
ack em
15:10:27 [Zakim]
em, you wanted to confirm the previously raised issue about contradictory information
15:10:31 [jjcscribe]
jjc: should we ask about charmod req
15:10:40 [jjcscribe]
that we are contravening
15:11:30 [DanC]
"management" includes the SemWeb CG, btw.
15:11:33 [jjcscribe]
em: is this information consistent with cannes?
15:11:40 [jjcscribe]
jjc: no it is not.
15:12:22 [jjcscribe]
ACTION em get W3c mgmt feedback on I18N issue
15:12:44 [jjcscribe]
ACTION jjc Dig out cannes minutes
15:13:20 [jjcscribe]
ACTION bwm give feedback to martin
15:14:27 [jjcscribe]
withdraw action bwm
15:14:32 [jjcscribe]
ACTION danc give feedback to martin
15:14:51 [jjcscribe]
end of I18N comment issue
15:15:00 [jjcscribe]
back to value space of XMLLiteral
15:15:09 [jjcscribe]
I18N issue seems to be parsing one
15:16:59 [jjcscribe]
jjc: discusses jang preference against octet sequences
15:17:23 [jjcscribe]
danc: qu are two bits of MXL equal is hard
15:17:39 [jjcscribe]
danc: webont want equality
15:18:57 [jjcscribe]
janG: every wg does its own thing on this
15:19:12 [jjcscribe]
jang: is XML equality technically hard, or procedurally hard
15:19:26 [jjcscribe]
danc: it is not hard, rather too easy, lots of answers
15:20:37 [jjcscribe]
jjc: we had the requitement to have equality, the specs of the consortium that had this were the c14n specs, so thats why we have this
15:22:17 [jjcscribe]
jang: the current value sapce is distasteful, but its not up to us to do better
15:22:55 [jjcscribe]
jjc proposes XML Literal value space
15:23:00 [gk]
The value space
15:23:01 [gk]
is the set of all exclusive Canonical XML (with comments, with empty
15:23:01 [gk]
InclusiveNamespaces PrefixList ), which when embedded within an arbitrary XML
15:23:01 [gk]
start and end element form a document conforming to XML Namespaces [XML-NS].
15:24:32 [gk]
"... canonicalized start and end element tags ..."?
15:24:41 [gk]
(That last my comment, not quite)
15:25:58 [DaveB]
DaveB has joined #rdfcore
15:26:16 [gk]
BWM: "The value space of XML literals is a set of exclusively canonicalized XML"
15:26:27 [jjcscribe]
jjc: clarifiies that editorial polishing may still be relevant
15:26:47 [jjcscribe]
jproposes gk lines
15:27:31 [jjcscribe]
proposed jjc, seconded jos, no abstentions, no agaionst
15:27:33 [jjcscribe]
resolved
15:28:05 [jjcscribe]
datatypes moved to email
15:28:10 [DanC]
actions arising from that decision? is there a commentor waiting? [I feel a bit lost; never mind if I am, indeed, lost]
15:28:18 [jjcscribe]
language tag or language identifier
15:28:45 [jjcscribe]
semantics uses tag, primer neither
15:29:00 [jjcscribe]
agendum 12
15:29:12 [gk]
DaveB, does syntax doc refer to labguage tag or identifier ??
15:29:20 [DaveB]
I don't recall
15:29:26 [gk]
;-)
15:29:55 [gk]
We think preferred term is "language tag" ...
15:30:11 [gk]
... for consistency acrosos docs. Do we care?
15:30:29 [jjcscribe]
editor has heard 'tag'
15:30:42 [jjcscribe]
schedule
15:30:50 [gk]
(RFC3066 uses "tag", and that's kind of the horse's mouth here)
15:31:05 [DaveB]
hmm, I find 'language identifier' once at least
15:31:25 [DanC]
re 14Jul... is that proposals from editors, or reviewed text for publication?
15:32:06 [DanC]
em, what form of "done" do you have in mind for 14Jul?
15:32:59 [DanC]
what's clear to me is that if we don't have a request for PR by 4 Aug, REC by ISWC is at risk.
15:33:05 [jjcscribe]
fm: primer aimed to be ready by 1 July
15:33:11 [jjcscribe]
jjc: concepts is ready
15:33:20 [em]
i think what i described last week.... all open issues resolved and docs at a level that they could be ready to publish for a 2nd last call (*not* suggesting this is required, but thought of this in terms of worse case schedule scenario)
15:33:20 [DaveB]
syntax is, or will be ready
15:33:25 [jjcscribe]
danbri: schema waiting on peter a bit
15:33:33 [DaveB]
(small items, I remembered a couple todo)
15:33:51 [jjcscribe]
semantics -
15:35:01 [jjcscribe]
bwm: which of PatH's version do we want?
15:35:19 [jjcscribe]
Gk: want entailement rules, but don't care about completeness claim
15:35:26 [jjcscribe]
jos, fm: +2
15:35:51 [jjcscribe]
danc: implementation experience?
15:36:11 [danbri_dna]
bwm, brian do you want to be muted?
15:36:20 [gk]
DanC, does implementation experience count for non-normative material?
15:36:26 [DanC]
I'm trying to work from the "mother may we have PR status?" perspective.
15:37:05 [jjcscribe]
bwm: I don't believe we need the completeness claim
15:39:38 [DanC]
ack danc
15:39:38 [Zakim]
DanC, you wanted to say I thought the rules were normative. I'm pretty sure lots of implementations are based on them. I think they actually do specify the interoperability of RDFS
15:39:41 [gk]
In support of jeremy: the benefit would be that we draw implementer feedback if we're wrong, so can move more quickly towards being truly complete. That said, I still don't care
15:42:55 [jjcscribe]
danc asks whether rules should be normative
15:43:13 [jjcscribe]
lots of people use the rules
15:43:29 [jjcscribe]
(connolly, jena, de roo .... lots of others)
15:43:49 [jjcscribe]
Ones that aren't: OWL Lite - DL implementations
15:45:03 [gk]
q+ to say I think this debate is not really important. To the extent that the closure rules are a logical consequence of normative semantics, they *are* normative.
15:45:08 [Zakim]
+Pat_Hayes
15:46:23 [jjcscribe]
Path speaks against normativity of closure rules
15:46:51 [jjcscribe]
they are valid it is a fact
15:46:59 [jjcscribe]
that they are complete is probably false
15:48:05 [gk]
q+ to argue against DanC's "effective decision procedure"
15:48:09 [jjcscribe]
q+ to ask pat about completeness
15:49:41 [jjcscribe]
PatH is not prepared to prove that RDFS is logically complete
15:50:22 [bwm]
ack gk
15:50:23 [Zakim]
gk, you wanted to say I think this debate is not really important. To the extent that the closure rules are a logical consequence of normative semantics, they *are* normative. and
15:50:25 [Zakim]
... to argue against DanC's "effective decision procedure"
15:50:43 [em]
q+ to balance picking our battles in the context of limited (and reducing) resources
15:51:19 [bwm]
ack danc
15:51:19 [Zakim]
DanC, you wanted to make the point about users expecations again
15:51:29 [bwm]
ack jjc
15:51:29 [Zakim]
jjcscribe, you wanted to ask pat about completeness
15:51:58 [bwm]
ack em
15:51:58 [Zakim]
em, you wanted to balance picking our battles in the context of limited (and reducing) resources
15:51:59 [DanC]
I thought the relevant theorem was "F1 rdfs-entails F2 iff datalog-closure(F1, rdfs-rules) rdf-simple-entails F2"
15:52:37 [bwm]
regrets from mike dean, patrick
15:53:44 [jjcscribe]
q+ to propose keeping editors conjecture as a compromise between normativity and dropping the rules
15:55:43 [jjcscribe]
danc: do we expect RDF RDFS reasoners to interoperate?
15:55:57 [bwm]
ack jjc
15:55:57 [Zakim]
jjcscribe, you wanted to propose keeping editors conjecture as a compromise between normativity and dropping the rules
15:55:59 [jjcscribe]
gk, path, bwm: we do not define reasoners etc
15:58:08 [jjcscribe]
jos was surprised to see rules move to an appendix
15:59:12 [jjcscribe]
danc: when there is a bug where rules and model theory differ
15:59:22 [jjcscribe]
danc: is this a bug in the rules or the mt?
15:59:33 [jjcscribe]
path: in the rules by defn
16:00:34 [jjcscribe]
gk: I don't want to be forced to use the rules for everything.
16:00:43 [jjcscribe]
q+ to talk about datatypes
16:01:18 [jjcscribe]
danc: suggests banana as an RDF implementation!
16:02:02 [DanC]
this tells me that RDFS is too complicated. I thought it was 12 horn rules.
16:03:06 [jjcscribe]
chair believes that it is late to raise normativity of rules
16:03:35 [em]
q+
16:04:39 [danbri_dna]
the wg is out of time. we have to finish. i'm happy moving rules to appendix if that's the only way to finish...
16:04:50 [bwm]
ack jjc
16:04:50 [Zakim]
jjcscribe, you wanted to talk about datatypes
16:05:10 [bwm]
ack em
16:06:32 [jjcscribe]
em speaks about how to get to PR
16:07:27 [jjcscribe]
bwm proposes closure rules remains in semantics doc and claim of completeness remains
16:08:22 [jjcscribe]
no formal decision
16:08:25 [DanC]
bwm, WG decisions are expensive. beware of saying "the WG must decide ..."
16:08:38 [jjcscribe]
pat offers to strengthen RDF case with a proof
16:08:52 [em]
i'm really sorry folks... i'm late and have to run. will lurk on irc for remainder of meeting
16:08:58 [Zakim]
-EMiller
16:09:19 [jjcscribe]
Meeting closes.
16:09:22 [Zakim]
-??P14
16:09:28 [Zakim]
-FrankM
16:09:42 [Zakim]
-GrahamKlyne
16:09:47 [Zakim]
-bwm
16:14:18 [gk]
DanC, to fly a kite, my story to the director would be that the entailment rules, though non-normative, to the extent that they are a logical cosequence of the normative semantics have normative force. Therefore there's no need for them to be normative.
16:31:59 [Zakim]
-ILRT
16:55:23 [Zakim]
-DanC
16:55:24 [Zakim]
-jjc
16:55:34 [Zakim]
-Pat_Hayes
16:55:35 [Zakim]
SW_RDFCore()10:00AM has ended
19:01:47 [Zakim]
Zakim has left #rdfcore
19:23:52 [DanC]
DanC has left #rdfcore