W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > September 2002

Minutes of RDF Core WG telcon, 2002-08-30

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Mon, 02 Sep 2002 11:38:56 +0100
To: w3c-rdfcore-wg@w3.org
Message-ID: <5090.1030963136@hoth.ilrt.bris.ac.uk>



Agenda of meeting:
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0266.html

IRC logs:
  http://ilrt.org/discovery/chatlogs/rdfcore/2002-08-30.html
  (attached)

1: Scribes

  Dave Beckett
  Jan Grant volunteered to scribe next week.

2: Roll Call

Participants:
  Dave Beckett (scribe)
  Dan Brickley
  Jeremy Carroll 
  Mike Dean
  Jan Grant
  Graham Klyne
  Frank Manola
  Brian McBride (chair)
  Eric Miller
  Steve Petschulat
  Patrick Stickler
  Aaron Swartz

Regrets:
  Dan Connolly 
  Jos De Roo
  Sergey Melnik

Apologies:
  Pat Hayes


3: Review Agenda

No AOB.


4: Next telecon 6th Sept 2002

Agreed.


5: Minutes of 2002-08-23 telecon with correction

See:
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0224.html

Approved.


6: Confirm Status of Completed Actions

All actions (see agenda) closed.


7: Status of new Concepts and Abstract Data Model WD

2002-08-16#3  EricM  publish rdf concepts and abstract data model WD

  EricM reported it was being processed by the publications team 

  Later on during the telcon it was published at
    http://www.w3.org/TR/rdf-concepts/
  and the group thanked Eric, Graham and Jeremy for the work in
  getting it published.

Closed.


8: Status of new mime types draft

2002-08-23#8  AaronSw  Update MIME type draft for WG review prior to re-issue.

Aaron noted he had been updating it immediately after receiving
comments.  Publication was approved, above action closed.

ACTION 2002-08-30#1 AaronSw: Publish new MIME Type Internet Draft.


9: Datatypes

Brian reviewed the progress made and started the discussion on part 1
of Patrick's new consensus building document:

  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0257.html

which Patrick introduced as intended to capture aspects of datatyping
with the emerging consensus, with part 2 as other aspects and
issues.  There was support for part 1 as a way forward and a
discussion of whether there was ever a goal for a minimal datatypes
solution.   Discussion of global datatyping in both 6.1.2 and 3.1
an dneed for distinguishing them and more examples.

Discussed rdf:type and DaveB's issue with the syntax:
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0260.html
which was found as a compelling example of a problem by Jeremy Carroll,
Dan Brickley.

ACTION 2002-08-30#2 Brian: Add an issue of the attribute used to
identify datatype literals.

Jan proposed to change the literals from a three part structure into
multiple literal forms and use such things as pairs of
(UnicodeString type, string) (LangString type, string, language)

ACTION 2002-08-30#3 Jan: Write a proposal to change the form of literals.

Brian offered that this seems to be a basis for moving forward
(although no comments from Pat Hayes or Sergey on this specific
document yet) and it seems that the group is happy to change RDF to
put datatypes in the abstract syntax, RDF/XML syntax and model,
primer, vocabulary description/schema.  In order to not specify
things in two places, maybe the datatypes material should be moved
into other document(s).

Jeremy noted that this has changed the groups positon from datatypes
being a layer on RDF(S), to being part of RDF(S) and this may have an
impact on document structure.


10: rdf:Alt's relationship to individual statements
2002-08-23#1  DanC  review primer text and propose alternative that clarifies
               status of Alt
Continued.

2002-08-23#2  DaveB  initiate dialog with DC community about the issue of
                rdf:Alt usage.
Done.  DaveB reported that it is hard to explain it to the DC since
there he has no full understanding of how to use rdf:Alt and would
like something to point them at.  Eric said that he thought DC saw
it as a collection mechanism, don't mind our choices as long as we
give clear guidance.

  dc-architecture thread on rdf:ALT
  http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0208&L=dc-architecture&T=0&F=&S=&P=3573


ACTION 2002-08-30#4 Brian: Try to enter the conversation on dc-architecture.


11: Relationship between XML Schema and RDF
Specifically, suggestion to ignore attributes from the xml:schema
namespace to allow processing by both xml schema processors and rdf
processors.
             
See
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0165.html


Skipped.


12: Meaning of rdf:type and relationship between RDF and RDFS

Frank agreed to remove the term "rdf processor"
See:
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0173.html

Discussion on that without a processor, there is no conclusion or
entailment so this need not be mentioned.  Discussion of if rdf:type
is in RDF, is rdfs:Class in another language, RDFS?   This is
described in the MT by the RDF/RDFS entailment rules.  The group
felt there was no need for decision here, as the MT defines the
answers to this area.


13: Frank's new assertion text
2002-08-23#7  FrankM  Propose alternative text for the concepts and abstract
                 model document to rectify concerns with conflicting use of 
                 "assertion".

See:
  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0226.html
  http://www.ninebynine.org/wip/RDF-basics/2002-08-05/Overview.htm#section-Social

Skipped.

------------------------
13:55:51 <Zakim> SW_RDFCore()10:00AM has now started
13:55:58 <Zakim> +FrankM
13:56:38 <Zakim> +??P1
13:56:47 <gk> zakim, ??p1 is gk
13:56:48 <Zakim> +Gk; got it
13:58:42 <danbri> danbri has joined #rdfcore
13:58:57 <Zakim> +??P2
13:59:07 <DaveB> Zakim, ??p2 is ILRT
13:59:08 <Zakim> +ILRT; got it
13:59:17 <DaveB> ILRT has bwm, daveb
13:59:23 <DaveB> Zakim: ILRT has bwm, daveb
13:59:30 <DaveB> phooey
13:59:35 <AaronSw> zakim, ilrt has bwm, daveb
13:59:37 <Zakim> +Bwm, Daveb; got it
14:00:06 <Zakim> +AaronSw
14:00:34 <em_lap> em_lap has joined #rdfcore
14:00:35 <DaveB> Zakim, ILRT has bwm, daveb, jang
14:00:37 <Zakim> Bwm was already listed in ILRT, DaveB
14:00:37 <Zakim> Daveb was already listed in ILRT, DaveB
14:00:38 <Zakim> +Jang; got it
14:00:52 <Zakim> +??P4
14:01:15 <AaronSw> zakim, ??p4 is SteveP
14:01:16 <Zakim> +SteveP; got it
14:01:26 <DanConn> DanConn has joined #rdfcore
14:01:28 <AaronSw> zakim, who's here?
14:01:29 <Zakim> On the phone I see FrankM, Gk, ILRT, AaronSw, SteveP
14:01:30 <Zakim> ILRT has Jang
14:01:31 <Zakim> On IRC I see DanConn, em_lap, danbri, gk, DaveB, bwm, Zakim, AaronSw, logger_1
14:01:36 <Zakim> +EricP
14:01:37 <danbri> Zakim, EricP is temporarily DanBri
14:01:38 <Zakim> +DanBri; got it
14:01:41 <AaronSw> Zakim, ILRT has bwm, daveb, jang
14:01:42 <Zakim> Jang was already listed in ILRT, AaronSw
14:01:43 <Zakim> +Bwm, Daveb; got it
14:01:47 <DaveB> DaveB is now known as db-scribe
14:01:49 <Zakim> +Mike_Dean
14:01:54 <AaronSw> zakim, ILRT also has JanG
14:01:55 <Zakim> +JanG; got it
14:02:01 <db-scribe> agenda http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0266.html
14:02:01 <AaronSw> hmm. zakim bug
14:02:03 <DanConn> em_lap, bwm, danbri, can you accept my regrets?
14:02:22 <AaronSw> bwm: ok
14:02:23 <danbri> ack'd
14:02:28 <Zakim> +EricM
14:02:32 <mdean> mdean has joined #rdfcore
14:02:49 <danbri> em, we got your voicemail
14:02:54 <em_lap> ack
14:03:12 <em_lap> zakim, please disconnect me
14:03:13 <Zakim> sorry, em_lap, I do not see a party named 'em'
14:03:20 <em_lap> zakim, please disconnect em_lap
14:03:21 <Zakim> sorry, em_lap, I do not see a party named 'em_lap'
14:03:25 <danbri> zakim, who is here?
14:03:26 <em_lap> zakim, please disconnect EricM
14:03:26 <Zakim> On the phone I see FrankM, Gk, ILRT, AaronSw, SteveP, DanBri, Mike_Dean, EricM
14:03:27 <Zakim> ILRT has Bwm, Daveb, JanG
14:03:28 <Zakim> EricM is being disconnected
14:03:29 <Zakim> On IRC I see mdean, DanConn, em_lap, danbri, gk, db-scribe, bwm, Zakim, AaronSw, logger_1
14:03:30 <Zakim> -EricM
14:03:38 <em_lap> * em_lap tries again
14:04:11 <db-scribe> item 1 scribe
14:04:16 <db-scribe> jang volunteers for next week
14:04:21 <db-scribe> item 2 roll call
14:04:26 <db-scribe> Zakim, who's here?
14:04:27 <Zakim> On the phone I see FrankM, Gk, ILRT, AaronSw, SteveP, DanBri, Mike_Dean
14:04:28 <Zakim> ILRT has Bwm, Daveb, JanG
14:04:29 <Zakim> On IRC I see mdean, DanConn, em_lap, danbri, gk, db-scribe, bwm, Zakim, AaronSw, logger_1
14:04:32 <danbri> zakim, mute danbri
14:04:33 <Zakim> DanBri should now be muted
14:04:39 <Zakim> +??P7
14:04:59 <AaronSw> zakim, ??P7 is jjc
14:05:00 <Zakim> +Jjc; got it
14:05:27 <db-scribe> regrets danc, jos
14:05:35 <db-scribe> regrets sergey
14:05:46 <DanConn> re my Alt action: no progress. apologies. pls continue.
14:05:53 <db-scribe> item 3 review agenda
14:06:11 <em_lap> * em_lap is awaiting dialout.... 
14:06:19 <DanConn> DanConn has left #rdfcore
14:06:22 <Zakim> +EricM
14:06:50 <em_lap> * em_lap remains mutted 
14:07:16 <db-scribe> bwm: apologies for agenda being late, didn't seem to mail out right
14:07:18 <db-scribe> item 4
14:07:23 <db-scribe> next telcon 6 sept 2002
14:07:33 <em_lap> re ericm action item on publishing... request is still being processed by pub team
14:07:36 <db-scribe> frankm: regrets for 6 sept
14:08:13 <db-scribe> item 5 minutes of 2002-08-23 telcon
14:08:22 <db-scribe> approved
14:08:30 <db-scribe> item 6 completed actions
14:08:34 <db-scribe> all done
14:08:39 <em_lap> yep
14:08:43 <db-scribe> item 7 concepts and abs data model wd
14:08:58 <db-scribe> em_lap: in pub team queue, hoped for today
14:09:26 <db-scribe> note to gk: check things for pubrules
14:09:29 <em_lap> http://www.w3.org/2002/08/05-rdf/
14:09:45 <db-scribe> item 8 status of mime types draft
14:10:31 <db-scribe> aaron: been updating after comments
14:10:34 <db-scribe> gk: time to goahead
14:10:39 <em_lap> gk, if you haven't started integrating any updates, please build on http://www.w3.org/2002/08/05-rdf/ this would help re future publications 
14:10:42 <danbri> zakim, unmute danbri
14:10:43 <Zakim> DanBri should no longer be muted
14:10:55 <db-scribe> danbri: happy don't wait
14:11:05 <db-scribe> ACTION aaron: publish new mime type ietf draft
14:11:10 <db-scribe> item 9 datatypes
14:11:16 <gk> My message: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0252.html
14:11:23 <danbri> zakim, mute danbri
14:11:24 <Zakim> DanBri should now be muted
14:11:35 <db-scribe> bwm reviews
14:11:37 <Zakim> +PatrickS
14:11:47 <danbri> rewind 15 seconds :)
14:12:18 <db-scribe> bwm: review of 2 parts of patricks' doc
14:12:27 <db-scribe> where part 1 captures the progress made, agreed on
14:12:37 <db-scribe> and part 2 is ... (other stuff)
14:12:56 <db-scribe> patricks: part2 is what the wg may choose to discuss
14:14:02 <db-scribe> gk: thought we were going for a minimal datatypes plan sufficient for numbers and other vals
14:14:38 <db-scribe> jjc: agree might be over strong; some sense of emerging consensus
14:15:30 <db-scribe> jang: part 1 a very good job in a neutral light of least contentious part.  May need more advice instead of options
14:17:18 <db-scribe> patricks: the two parts mean we can see way to possibly select bits of part2 above the minimal solution in part1
14:18:24 <db-scribe> bwm: part1 review pls
14:19:51 <db-scribe> patricks: can we leave off the rdf:type discussion to the list, where examplesa rew beter discussed
14:19:59 <db-scribe> jjc: I find daveb's example compelling
14:20:11 <db-scribe> frankm: can we leave discussion of what attr, over if there is an attr at all
14:20:38 <db-scribe> bwm: records issue of the attr used to identify the datatype
14:20:59 <danbri> fwiw, as per jjc's comment, I also found dave's argument compelling.
14:21:25 <db-scribe> frankm: use of the word constraint in sec 3.1 might be trouble, best used without that word (just assertions)
14:21:34 <db-scribe> patricks: can try to re-rexpress
14:22:09 <danbri> [background noise]
14:22:11 <db-scribe> frankm: problem with 6.1.2 and see 3.1's global datatyping
14:22:49 <db-scribe> 6.1.2 excluding the use of inline literals with the range constraint properties
14:22:59 <db-scribe> and comparing with 3.1 is confusing
14:23:18 <db-scribe> no instance example in 3.1 of somebody specifying an age
14:23:39 <db-scribe> how do you specify an actual age in an instance?
14:24:09 <db-scribe> patricks: global implicit dataypes uses the range assertion to assign a datatypes that is not other specified
14:24:17 <db-scribe> (not in part 1)
14:24:36 <db-scribe> if we have explict locally typed literals, then the machinery for rdf:range works as it does wnow
14:24:46 <db-scribe> and we can assert the type using them
14:25:31 <db-scribe> frankm: so, in 3.1 this would be ok in an instance that explictly gives the datatype.  Would be clarified with an example.
14:25:40 <db-scribe> patricks: yeah, need an example in 3.1
14:25:52 <db-scribe> frankm: in 6.2, need to compare with the (new) example in 3.1
14:26:03 <db-scribe> patricks: ok
14:26:50 <db-scribe> stevep: unclear when the document creator has to duplicate the types
14:28:13 <db-scribe> frankm: 6.2 the property cannot validly accept inline literal values
14:28:33 <db-scribe> patricks: need to say when MT provides no interpretation for some things
14:29:09 <db-scribe> bwm: editorial things?  any more substantial issues?
14:30:04 <db-scribe> jjc: graham talked of a minimal spec. but part 1 has a range constraint that might be removed?
14:30:21 <db-scribe> gk: had a quick look, most of my previous comments are applicable
14:30:59 <danbri> gk's comments (just posted): http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0269.html
14:31:00 <db-scribe> patricks: not sure if we are going for an absolute minimal spec.  We didn't decide that
14:33:01 <danbri> zakim, unmute danbri
14:33:02 <Zakim> DanBri should no longer be muted
14:33:19 <db-scribe> bwm asks and gets conf that the spec need not be minimal
14:33:35 <db-scribe> frankm: really happy with part1 as far as it goes
14:33:37 <db-scribe> solid
14:34:03 <danbri> danbri: i would like to see _something_ rather than _nothing_; we need to do the range thing eventually, don't know if we can get there in this step
14:34:08 <db-scribe> bwm: gk, do a proper read through and review
14:34:49 <db-scribe> jjc: minor issue with xml union datatypes interaction of range constraints.  May not be difficult
14:35:26 <db-scribe> patricks: went over with patH earlier, there isn't a problem.  It satisfies the three characteristics of an rdf datatype
14:35:57 <db-scribe> jjc: editorial issue that is sufficiently non-obvious that should be recorded
14:36:16 <db-scribe> jang: expressing a literal as a dataype + language + xml bit - like to de-emphasis this
14:36:18 <db-scribe> NOISE
14:36:19 <danbri> [horrible white noise]
14:36:21 <em_lap> ouch?
14:36:26 <danbri> zakim, who is speaking?
14:36:27 <Zakim> I don't understand your question, danbri.
14:36:32 <Zakim> +AaronSw.a
14:36:33 <danbri> zakim, who is talking?
14:36:38 <em_lap> zakim, who is talking
14:36:39 <Zakim> I don't understand 'who is talking', em_lap.  Try /msg Zakim help
14:36:47 <Zakim> danbri, listening for 10 seconds I heard sound from the following: FrankM (100%), Gk (41%), AaronSw (100%), DanBri (45%), AaronSw.a (14%)
14:36:50 <db-scribe> silence
14:36:52 <Zakim> -AaronSw
14:37:08 <db-scribe> jjc continues - de-emphasis the 3-part structure
14:37:16 <db-scribe> typed literals which are unicode strings
14:37:23 <db-scribe> typed literals which are xml
14:37:31 <db-scribe> see current syntax as sugar for them
14:37:40 <danbri> zakim, who is muted?
14:37:41 <Zakim> I see Jjc muted
14:37:42 <db-scribe> bwm: how do we represent literals in the abstract literal, including datatypes
14:38:02 <db-scribe> jang: change to the mindset - type is UnicodeString, other bit is a uncide string
14:38:10 <db-scribe> type is LangString, other bit is lang+unicode string
14:38:21 <db-scribe> ACTION jang: write up a proposal for this
14:39:00 <db-scribe> patricks: not sure if this is necessary for datatypes?
14:39:11 <db-scribe> jang: yes, structure is not important
14:39:48 <db-scribe> patricks: xml:lang inffects all contained XML
14:39:54 <db-scribe> bwm: can we take offline and ask parser writers
14:40:00 <db-scribe> bwm: can we talk about rdf:type
14:40:39 <db-scribe> bwm: I see the experts in the group all are speaking with one voice
14:40:54 <gk> DaveB gave a concrete example
14:41:29 <db-scribe> jjc: rdf:type makes properties vs something that doesn't give an arc
14:42:14 <gk> Dave's example of bad rdf:type behaviour is in http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0260.html
14:42:41 <db-scribe> bwm: this is the basis for moving forward?
14:43:10 <db-scribe> bwm: no comments or input yet from PatH or sergey on this specific doc
14:43:29 <db-scribe> how does this datatyping doc relate to the other specs
14:43:41 <db-scribe> seems we are changing the abs syn of RDF - reflect in concepts and abs dM doc
14:43:52 <db-scribe> changing (rdf/xml) syntax in order to indicate dataype of literal
14:44:06 <db-scribe> and schema documetn will have to address domain and range on datatypes
14:44:20 <db-scribe> don't want to specify things in 2 places
14:44:22 <Zakim> -SteveP
14:44:37 <db-scribe> bwm: dataypes editors, think about this
14:44:54 <danbri> Just wanted to note: the RSS group are looking at using range constraints on subclass of literals, see http://groups.yahoo.com/group/rss-dev/message/3264 
14:45:25 <db-scribe> patricks: don't see any dependency on rdfs, no change to rdfs:range, doesn't use rdfs:domain
14:46:00 <gk> q+ to agree that RDF schema not affected, but abstract syntax and XML syntax may be
14:46:01 <Zakim> * Zakim sees Gk on the speaker queue
14:46:29 <bwm> ack zakim
14:46:31 <Zakim> * Zakim sees Gk on the speaker queue
14:46:36 <bwm> ack Gk
14:46:37 <Zakim> Gk, you wanted to agree that RDF schema not affected, but abstract syntax and XML syntax may be
14:46:38 <Zakim> * Zakim sees no one on the speaker queue
14:47:21 <db-scribe> frankm: datatyping may not change the specifications but it may change what's said in, say, the schema document
14:49:17 <db-scribe> frankm: I'll want to say in theschema section of the primer about datatypes used
14:49:33 <db-scribe> [not sure I captured that right]
14:50:12 <db-scribe> jjc: we conceived this as a layer; once we start putting this into a graph, can't keep the layering view - impact on doc structure
14:50:22 <danbri> (off telecon to save time: as RDFS editor, I do expect that spec will need to field expectations re datatypes. I'm tempted to add a bookPrice example to the current eg)
14:50:49 <em_lap> hear hear!!
14:50:51 <db-scribe> bwm: we have made a lot of progress (lots, while I was away)  - thanks a lot patrick
14:50:55 <Zakim> -AaronSw.a
14:50:59 <db-scribe> bwm: need to maintain momentum
14:51:40 <db-scribe> item 10
14:52:47 <db-scribe> danc action continued
14:52:50 <db-scribe> daveb done
14:52:51 <em_lap> q+
14:52:52 <Zakim> * Zakim sees Em on the speaker queue
14:52:57 <gk> Dave has spoken to DC group, but cannot answer their questions because he doesn't fully understand himself how to answert theior questions
14:52:58 <em_lap> q-
14:52:59 <Zakim> * Zakim sees no one on the speaker queue
14:53:04 <db-scribe> action bwm: try to enter the conversation
14:53:21 <danbri> aside: Dublin Core architecture list, http://groups.yahoo.com/group/rss-dev/message/3264
14:53:41 <danbri> alt thread: http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0208&L=dc-architecture&T=0&F=&S=&P=3573
14:54:51 <db-scribe> em: dc groups seing this as a collection mechanism
14:55:28 <db-scribe> em: don't think dc group would mind one way or other, need guidance.  waiting for us
14:56:22 <gk> From last week, IIRC, note that one concern with rdf:Alt previously noted was that it might be used with existing properties
14:57:08 <db-scribe> item 12 meaning of rdf:type rdf/rdfs relationship
14:57:27 <db-scribe> bwm: removing "rdf processor"?
14:57:58 <em_lap> q+
14:57:59 <Zakim> * Zakim sees Em on the speaker queue
14:57:59 <db-scribe> frankm: yes, will remove such refs but still would like a agreement on bwm'sentailment
14:58:11 <em_lap> q+, to mention http://www.w3.org/TR/rdf-concepts/ 
14:58:12 <Zakim> * Zakim sees Em on the speaker queue
14:58:29 <db-scribe> frankm: suggest that an (rdf processor) is not required to conclude that the object of a type property is  a class
14:58:37 <db-scribe> jjc: that is an entailment
14:59:31 <db-scribe> danbri: no concluding (processing entailment rules) if you don't do entailment
14:59:34 <danbri> it seems too lose to me; "an rdf processor is not required to conclude ANYTHING at all"... 
14:59:57 <db-scribe> frankm: rdf:type is in RDF, rdfs:Class is not?
15:00:22 <db-scribe> frankm: what concepts are one lang, what are in another?
15:00:35 <db-scribe> jjc: the MT clarifies this
15:01:55 <db-scribe> bwm: no need for a decision; this is implicit in the MT?
15:01:58 <db-scribe> frankm: ok
15:02:37 <Zakim> -PatrickS
15:02:38 <Zakim> -DanBri
15:02:41 <db-scribe> thanks to gk and jjc for new WD: http://www.w3.org/TR/rdf-concepts/
15:02:43 <Zakim> -Mike_Dean
15:02:45 <db-scribe> end of meeting
Received on Monday, 2 September 2002 06:41:59 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:50:55 EDT