2001-07-27 minutes

Minutes for 2001-07-27 below. I'm terribly sorry for the delay. I was
working from a rushed 1st pass at them which I tried to do in aeroplanes
and hotels prior to the F2F. Scrapping that and starting over proved
quicker. --danbri


RDFCore WG 2001-07-27 Teleconference Minutes






Roll call:


- Jos De Roo
- Bill DehOra
- Martin Horner
- Ora Lassila
- Frank Manola
- Aaron Swartz
- Mike Dean
- Pat Hayes
- Dan Brickley (scribe)
- Art Barstow
- Brian McBride (chair)
- Eric Miller

regrets from:
 - Dave Beckett
 - Dan Connolly
 - Jan Grant
 - Ron Daniel
 - KWON, Hyung-Jin
 - Yoshiyuki Kitahara
 - Satoshi Nakamura

 - Sergey Melnik
 - R.V.Guha
 - Frank Boumphrey
 - Rael Dornfest
 - Graham Klyne
 - Michael Kopchenov
 - Stephen Petschulat
 - Pierre G. Richard


Review of Previous Action Items:

 Previous minutes:



        - Renato has resigned from the WG
 	- a correction proposed for character-encoding issue
	  [action] on em to write up and send to list
	- WG accepted Brickley's recommended changes to M+S re containers
	- re Bill's msg re his ascribed [action], bill to send mail clarifying his
	  action regarding review of M+S w.r.t. what it says about literals.

Minutes of last week accepted with these corrections.


Confirm Following Actions Completed

ACTION 2001-07-13#3 Frank Manola: Draft an assessment of Coordination
    points between RDF(S) and DAML+OIL

ACTION 2001-07-20#3 - Brian - write up Pat's notes and sent it to
        the list

ACTION 2001-07-20#4 - all WG members - send references to model
        theory proposals to the list

ACTION 2000-07-20#7 - Graham - send additional thoughts/comments
  to the list.  Include questions [and answers from Graham's perspective]
  that will help clarify the issues.

ACTION 2000-07-20#9 - Graham - create xml:lang use cases

ACTION 2000-07-20#11 - Brian - propse changes that need to be made to
        clarify the M&S

ACTION 2000-07-20#12 - FrankM - send model theory refernces to the list

AOB / Agenda review:

  additions - Frank's review of DAML+OIL
            - Pat's model theory;  both arrived since agenda sent out.

  No AOB



ACTION: 2001-06-22#5: DanBri: Get a draft of RDFSchema to the group.

ACTION: 2001-07-06#2 (danbri) - write test cases for the RDF schema issues
   that are impacted by literals-as-resources.

ACTION 2001-07-13#8 Pat Hayes: Produce a strawman model theory for RDF

ACTION 2001-07-13#10 Pat Hayes: Provide his point of view on anonymous
DROPPED (included in model theory)

ACTION 2001-07-13#11 Graham Klyne: Summarize the proposals on anonymous
  resources from Frank, Pat, and any others

ACTION 2001-07-20#1 - all WG members - send regrets to Eric Miller ASAP
       if you not be able to attend the meeting.

ACTION 2001-07-20#2 - all WG members - if you registered but are not
       on the registration list:


       send email to Eric Miller.

ACTION 2001-07-20#5 - Sergey - summarize the model theory references

ACTION 2001-07-20#6 - DanBri - send a note to the list that
        describes the RDF Schema issues that he wants to discuss
        at the f2f meeting

ACTION 2000-07-20#7 - Graham - send additional thoughts/comments
  to the list.  Include questions [and answers from Graham's perspective]
  that will help clarify the issues.

ACTION 2000-07-20#8 - Frank - send additional thoughts/comments on anon
  resources to the list

ACTION 2000-07-20#10 - Bill - create xml:lang use cases

----- Agenda Items -----

Summary of new actions arising from Agenda Items:

ACTION 2001-07-27#1 - EricM - For 2001-07-20 minutes re character-encoding
	issue, to write up his correction and  send to list

ACTION 2001-07-27#2 - Bill - For 2001-07-20 minutes, Bill to send mail
	clarifying his action to review M+S w.r.t. what it says about literals.

ACTION 2001-07-27#3 - Brian - Update f2f page to point to Pat's model theory

ACTION 2001-07-27#4 - Pat - create table of pros/cons for model theory for f2f

Discussion details....


Preparing for the f2f (10 mins):

Review the updated F2F page at:


especially objectives, agenda and required/recommended reading.

ACTION 2001-07-27#3 - Brian - Update f2f page to point to Pat's model theory

Review draft response to DAML+OIL Joint Committee:


 - detailed discussion was postponed until after the F2F

Issue rdfms-xmllang (5 mins):


Review proposal at:



 brian: any comments on my note?
 martin: of 3 test cases, mine showed no interest in
      xml:lang issue. We found it wasn't a major block to implementation.
      This is reflected fairly in Brian's summary.

 brian: I propose "xml lang as defined in m+s is useful. would be wrong to
	change ntriple and model theory until we've considered
 (general approval, though noted that Sergey was not present)

Review status of #rdfms-identity-anon-resources (10 mins)


	"This is a key issue for the f2f.  We need to have a framework to
	approach it."

brian: we've seen 160+ msgs. Where do we stand on this?

jos: In my opinion, we (I?) have clear view on this, from Pat's doc.
	 ... pat has collected all the different positions and
	unified them in one nice document. Perhaps not final but at least tangible.
	The StrawDog doc of yesterday
	(ie. the first revised draft of the model theory doc,
	http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0399.html )

brian: is that the general view?
pat: Frank's msg summarised a number of alterantives pretty usefully

frank: 	two sub-threads here. Getting straight what the interpretation of
	anonymous nodes will be. And also deciding what to  do whether have URIs or

Jos: we at least have description of the tradeoffs now

pat: what the model theory does, is make the alternatives clearer, but
	doesn't decide between them.
	...someone has to decide which way to go. Either way makes sense; i've no
	strong feeling.

brian: the thing worrying me. I can't see a framework for how we make that

pat: i volunteer in time for my f2f presentation, to draw up list of pros
and cons of the two main alternatives and their utility. Everything I can
think of in a little table.

ACTION 2001-07-27#4 - Pat - create table of pros/cons for model theory for f2f

Review status of #rdfms-uri-substructure (10 mins):

brian: sergey proposes resources should be identified by pair of strings not URIs
 I feel torn by this; I'm drawn to it, but feels like a fundamental shift
from Web Architecture. Using something other than URis to identify things beyond our scope?

Sergey's proposal: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0270.html


Review status of #rdfms-literal-is-xml-structure (10 mins):

bill: i'd like to send a big doc to list shortly
	...not to sure how far we can get by sying parseType=Literal parses to
	Infoset until we know what existing M+S says. I quite like the infoset
	proposal but not sure if that's in scope.

aaron: ...even something that could be converted into an infoset, eg.
	something with namespaces etc attached so the needed info is captured

bill: ...section 2 suggests we simply treat literals as strings. then you
	get into M+S and realise literals implicitly a structured object, xml:lang
	etc, language like "is part of the literal"
	...I won't be at f2f, would like to see discussion of literals as infoset. I
	think probably a godo thing

pat: i agree, merely being strings is too restrictive

jos: agree, recommend people look at TimBL's design issues on Interpretation

bill: if you take RDF's simplest "hello world" Ora example, _what does that
	mean_ ?
	...a literal is something that denotes or evaluates to itself? eg. if you
	have foo:creator = <some infoset>, do we mean the infoset created it? Or
	rely on the application semantics?
	...is the string itself the creator versus the thing denoted by the string

Review status of #rdfms-graph (10 mins):

responses to Pat's Model Theory:

pat: a question w.r.t Ntriple. Is the sense of the group that Ntriple will be offered
	to world as a sort of standard notation for rdf? or just for us? or

bill: last week i think we said it'd go in as an appendix

brian: ntriple started off as a way of doing test cases; a concrete syntax
	for representing triples for machine processing, comparing results etc
	...i've been suggesting / asking the question, can we make nntriple like the
	core of all these specs; defining syntax in terms of transform into ntriple,
	and model theory over ntriples?

pat: makes sense to me; i'm from a tradition that prefres textual language

brian: we're not proposing this to the world as a mechanism for
interchanging RDF

pat: ok, makes sense

brian: we're not doing it as "a new syntax"; enough to do, already running
	late so don't want to do things beyond charter

aaron: but if we happen to do one "in passing", nobody'll mind?

[some discussion of URIs and notion of meaning of a URI being common across
all interpretations?]

pat: issue is whether interpretation works over all universes
	...danger of using up all names at once
	names exist for a single interpretation

brian: one does lose something... can we get it back?
	...without going the whole way...
can we represent the relationship between interpetations?

brian: (to pat) Is there way of representating the specialness of URIs, ie
that all interpreations need to treat URIs as constrained w.r.t. what they
denote, without making the grand claim that there is just one

pat: only thing currently is that literals mean same thing
in all interpretations. All other URIs are open; mean one thing in one
interpretation, other in others.

frank: do you distinguish between web uris and other uris

pat: no

danbri: that seems right to me. No fast line between 'web' and real world
	resources. Books, java classes, telephones etc

frank: things named 'on the web' seem at least to have some constant
	interpretation. Other uses of URIs for non webby things don't seem so

pat: one way to talk about this is to talk about two theories, documents,
	sets of rdf, being or sharing ...
	...the web would be some shared core of interpretation; one big world-wide
	interpretation. Those URIs that have a common meaning that everyone accepts.

aaron: reminds me of reminds me of http://www.w3.org/Architecture/state

frank: there's a belief that if its on the Web, everybody shares it

pat: I know URL is no longer an encouraged term, but the URIs machines use
	for getting web pages... have a fixed interpretation

Next meeting - 10am PST, 1st August 2001, Sebastopol

Next telecon - 10am Boston time, 10th August 2001. [cancelled due to power outage]


Received on Thursday, 16 August 2001 10:40:21 UTC