Minutes: JUN-01-2001 WG Teleconference

RDFCore WG 2001-06-01 Teleconference Minutes



 ACTION: <action_id>: <owner>: <text> - a new action item, where 
   action_id is an identifier for the action item, owner is the name 
   of the person assigned the action and text is the action item.

 RESOLUTION: <text> - a WG decision





Roll call:

  - Art Barstow (scribe)
  - Dave Beckett
  - Dan Brickley (chair)
  - Brian McBride (chair)
  - Dan Connolly
  - Jos De Roo
  - Bill dehOra
  - Mike Dean
  - Jan Grant
  - Martyn Horner
  - Ghram Klyne
  - Ora Lassila
  - Frank Manola
  - Eric Miller
  - Stephen Petschulat
  - Aaron Swartz

  - Frank Boumphrey
  - Rael Dornfest

  - Ron Daniel
  - Renato Iannella
  - Yoshiyuki Kitahara
  - Michael Kopchenov
  - Satoshi Nakamura
  - Pierre G. Richard

Review of Previous Action Items:

 Previous minutes:


D1: Dave Beckett  assemble test cases re #rdf-ns-prefix-confusion
                   and suggest details of how the grammar in the spec
                   should be updated.

A1: Dan Brickley  Solicit RDF feature usage info from Guha and report back to
                   the group.

A2  Dan Brickley: send analysis (#rdf-container-syntax-ambiguity
                   and #rdf-containers-syntax-vs-schema)to rdfcore-wg list 
A3: Dan Brickley  send analysis (#rdf-container-syntax-ambiguity and
                   #rdf-containers-syntax-vs-schema)to rdfcore-wg list
A4: Guha:         re #rdfms-reification-required: Present analaysis to list
                   for discussion.
A5: Brian McBride Link test cases, results etc. from issues list
A6: Jan Grant     Do an analysis of the impact of XML Base and summarise to
  DONE [discussion follow during other agenda items].
A7: Brian McBride Contact Rael about hosting face to face at O'Reilly.
A8: Graham Klyne  to summarize www-rdf-logic perspective of reification as
                   it applies to both logic and rdf andreport back to rdfcore wg

A9: Brian McBride edit the errata per the resolutions above; i.e. those
  CONTINUED. [ArtB to work with Brian to get Brian write access to
    to the errata and to determine a strategy for maintaining the
    appropriate documents.]

A10: Jan Grant    propose expected results format.
  DONE [discussion follow during other agenda items].


New Action Items [assigned during Review of Previous Action Items]

ACTION: JUN-01-01-#1: Martyn: create test cases for a previously assigned

ACTION: JUN-01-01-#2: FrankM: create tests case for action item A8.


Issues Discussion

Issue #1
   Owner Dan Brickley
   Discussion re understanding of the issue as in Dan/Ora's message:

   A further message is expected prior to the call.


Issue #2
   Owner: Jan Grant
   Resolve disposition of test cases in:

 Jan: I consider this done and want feedback or it will go into
   the errata.

 ACTION: JUN-01-01-#3: Jos: review the test cases

 ACTION: JUN-01-01-#4: DaveB: review the test cases

Issue #3


   Owner: Stephen Petschulat

   Discuss Stephen's proposal in:


 RESOLUTION: whereas the WG sees insufficient implementation experience 
             of the aboutEachPrefix feature in RDF, it does not belong 
             in a W3C Recommendation.


Any Other Business

AOB #1 - Test case format

 [Ed note: see the IRC log for more information.]

 Jan proposed a test case format in:


 DanC proposed an alternate format in:


 Jan stated that he thought DanC's proposal was OK.  No one else objected
 to Dan's format

 ACTION: JUN-01-01-#5: Eric: investigate a test case repository for the
   WG.  The repository must facilitat WG members adding and modifying
   test cases.

 ACTION: JUN-01-01-#6: DaveB: send a note to the WG that contains
   a pointer to the test case description schema and the RDF parser
   tests site.
AOB #2 - XML base and RDF

 [Ed note: see the IRC log for more information.]

 Jan: XML Base as a relatively simple spec that gives a way 
 to attach a base URI to attributes and element content but the spec
 doesn't say what to do with it.  It provides a way to anchor relative
 URIs when they occur in the serialization.

 DanC: I think it could be useful but without real implementation
 experience I think it should be added to the "some day pile".

 ACTION: JUN-01-01-#7: Jan: summarize what XML might do for RDF. 

AOB #3 - aboutEach

 [Ed note: see the IRC log for more information.]

 The basic questions are: how is aboutEach being used today; if 
 it is not useful, should it be removed from the spec?

 DanC: my short answer is that it should be removed from the core RDF
 and if it is indeed useful it could be layered on top of the core.

 DaveB: I agree with DanC that it should be layered.  It causes
 implementation [look ahead] problems when dealing with large
 RDF documents.

 Ora: I use it and do not want it removed.

 Brian: in some discussions with DanBri regarding aboutEach, we felt
 it was useful.

 ACTION: JUN-01-01-#8: DanC: send a note to the WG that points
   out the bug in the spec regarding aboutEach

 ACTION: JUN-01-01-#9: Ora: send a note to the WG that describes
   how aboutEach is being used and how it was implemented

AOB #4 - Meeting registration

 DanC: I want to inform the chairs that the W3C's meeting 
   registration system could be used for the WG's f2f meeting

AOB #5 - Determing a name for the test [output] format name

 RESOLUTION: the name will be n3-triples


Next meeting

 June 8, 2001, 10:00am Boston time

 Scribe: Stephen


IRC log

09:59:07 <dajobe-jang>	+SteveP
09:59:13 <dajobe-jang>	+MikeD
09:59:25 <ArtB>	+ArtB
09:59:34 <ArtB>	+Martyn
09:59:34 <dajobe-jang>	+martyn
09:59:40 <dajobe-jang>	:-)
09:59:46 ---	You are now known as scribe
09:59:48 -->	AaronSw (aswartz@ has joined #rdfcore
09:59:58 <scribe>	+EricM
10:00:24 -->	mdean (mdean@hh1114013.direcpc.com) has joined #rdfcore
10:00:34 ---	mdean is now known as _mdean
10:01:33 <scribe>	+Ora
10:01:50 <danbri>	i'm dialing  +1 630 536 3003 and losing :(
10:01:51 <AaronSw>	Aaron Swartz
10:01:53 <GK>	Graham Klyne
10:01:54 <_mdean>	+Mike Dean
10:01:55 <dajobe-jang>	Dave Beckett, Jan Grant
10:01:56 <scribe>	Art Barstow
10:02:17 <em>	+eric miller
10:02:22 -->	DanC (connolly@adsl-208-190-202-206.dsl.kscymo.swbell.net) has joined #rdfcore
10:02:33 <scribe>	Present: Ora, FrankM, Martyn, Jos, StephenP 
10:02:42 <bwm>	http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001May/0283.html
10:02:55 <AaronSw>	rael sends his regrets -- he is unable to attend for personal reasons
10:03:00 <danbri>	can someone confirm http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001May/0283.html -> +1 630 536 3003   
10:03:05 <scribe>	Regrets: Rael, Bill
10:03:07 <AaronSw>	confirmed
10:03:22 <em>	danbri: are you getting in to bridge but not able to join conference room, or not able to get answer to number at all
10:03:24 <scribe>	Agenda: test repository, XML base
10:03:36 <danbri>	no answer, bad tone. 
10:03:42 <em>	danbri, yes this is correct +1 630 536 3003 room #3003
10:03:53 <scribe>	Regrets: FrankB
10:03:53 <danbri>	don't wait for me. 
10:03:55 *	danbri  fiddles
10:04:01 <scribe>	Regrets: DanC
10:04:17 <dajobe-jang>	danbri: no # in room number
10:04:36 <em>	so you can't get answer calling +1 630 536 3003
10:04:39 <scribe>	Action D1: done
10:04:41 <danbri>	sure; haven't got that fare
10:04:54 <danbri>	My actions need to carry over to next week.
10:05:07 <scribe>	A1: Dan Brickley: need to carry over
10:05:14 <DanC>	phtpht
10:05:18 <scribe>	A2 and A3: Dan Bri carry over
10:05:29 <scribe>	A4: Guha: carry over
10:06:05 <scribe>	A5: Brian: have done some test case linking (DaveB's, Jan's, DanBri's)
10:06:13 <scribe>	... DanBri - reification
10:06:19 <scribe>	... DaveB - ns prefix confusion
10:06:28 <DanC>	scribe, DanC on the phone now.
10:06:46 <scribe>	Martyn: I will create some test cases for a past action item
10:06:47 *	AaronSw  waves
10:06:48 <danbri>	Guha and I spoke late yesterday; we (I) need to write that up. Continue the action.
10:07:04 <scribe>	A6: Jan: I sent a note to the list
10:07:07 <scribe>	... Done
10:07:30 <scribe>	A7: Brian: f2f is on, Aug 1 and 2
10:07:30 -->	dehora (mirc@modem-231.chromis.dialup.pol.co.uk) has joined #rdfcore
10:07:50 <scribe>	... Sebastibol - 60-90 minutes North of SFO
10:08:02 <DanC>	order... this is review of actions, not discussion.
10:08:08 <scribe>	... Done
10:08:13 <scribe>	A5: Done
10:09:01 <scribe>	A8: Graham [and FrankM]: I posted a message about this to the list; I just posted an update this morning
10:09:21 <DanC>	folks, in the interest of time, when your action comes up, please be prepared to answer "done." or "please continue this" or "I'd like to withdraw this."
10:09:24 <scribe>	... FrankM did a good job; if group concurs, we can call it Done
10:10:42 <scribe>	FrankM: in a response to a response from DanBri, I promised to create some test cases; the test cases are partly done; there will be more test cases than I would have thought
10:10:44 <DanC>	I guess there's another action review response option: "please put this on the agenda".
10:11:25 <danbri>	[danbri: My office phone appears to have had its international dialing privileges removed by Uni of Bristol beauracracy; *danbri fumes uselessly]
10:11:55 <scribe>	ArtB: do we mark A8 done and create a new item for FrankM's tests?
10:11:59 <scribe>	A8: Done
10:12:07 <em>	danbri, can you take international calls?
10:12:10 <scribe>	ACTION: FrankM: create the test cases.
10:12:14 <em>	can we call you?
10:12:25 <scribe>	Brain: I still do not have access to the errat
10:12:44 <em>	e.g. can the bristol folk call you and patch you in?
10:12:46 <scribe>	A9: have a new section in the issues list at the top
10:12:58 <scribe>	... the errata now has a pointer to the issues list
10:13:24 <scribe>	Brian: I am going to work with ArtB about how to structure the docs
10:13:34 <scribe>	... A9 = ongoing
10:13:48 <scribe>	A10: Jan: done [will revisit later]
10:14:03 <danbri>	re my phone: don't let this eat up your attention. em, yes I can take calls. I've asked someone here to chase it, but won't be immediate.
10:14:11 <scribe>	Issue 1: Dan Bri - will skip
10:14:29 <scribe>	Issue 2: empty property elements
10:14:29 <em>	danbri, you're on the agenda so it seems having you on the call would be a good thing :)
10:14:57 <scribe>	Jan: get back to me if you have any issues; otherwise, it will go in the errata
10:15:17 <danbri>	I concur; though talked with brian just earlier about reprioritising the agenda due to lack of progress on that issue. I'm trying to get dialed in tho.
10:15:20 <scribe>	Jos: I checked it but need to look again
10:15:40 <scribe>	DaveB: I also did a first glance and they look OK; want to do a detailed analysis
10:16:05 <scribe>	em: who will take an Action item to review the tests?
10:16:17 <scribe>	ACTION: Jos: review the test cases.
10:16:35 <scribe>	ACTION: DaveB: review the test cases
10:16:36 <DanC>	jan, the test are in your message of Fri, 1 Jun 2001 10:50:21 +0100 (BST) , right?
10:16:55 <dajobe-jang>	http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jun/0000.html
10:17:03 <scribe>	Agenda #3: Stephen
10:17:11 <scribe>	... aboutEachPrefix
10:17:25 <scribe>	... people generally want it to go - it is broken
10:17:38 <scribe>	... there are various options
10:17:52 <scribe>	... can deprecate and recommend a diff approach - this is my pref
10:17:58 <scribe>	Ora: I want it to be gone
10:18:14 <scribe>	Graham: agree with Ora, it should be layered
10:18:24 <scribe>	+1: Aaron, Jos
10:18:43 <scribe>	Jan: can we inherit this functinaly from somewhere else?
10:18:44 <dehora>	agree with Ora
10:18:48 <DanC>	am I being heard?
10:18:53 <em>	no
10:19:04 <em>	we can not hear you on the phone
10:19:04 -->	martynh (chatzilla@ANice-101-2-1-161.abo.wanadoo.fr) has joined #rdfcore
10:19:18 <scribe>	+DanC
10:19:37 <scribe>	DanC: no, cannot inherit it
10:20:04 <scribe>	DanC: I agree to remove it; does anyone use it?
10:20:21 <scribe>	AaronSw: SiRPAC [online] has some support
10:20:34 <scribe>	em: PICS uses it; 
10:20:45 <scribe>	... but containers would be a better way 
10:21:08 <scribe>	DanC: perhaps this goes in a someday pile
10:21:31 <scribe>	... that is consider this for later work
10:21:35 <scribe>	Ora: good idea
10:21:54 <scribe>	Brian: Art and I will consider how to doc this
10:22:15 <scribe>	Brian: DECISON: aboutEach* will be removed from the spec 
10:22:32 <dajobe-jang>	NO
10:22:38 <dajobe-jang>	aboutEachPrefix
10:22:55 <scribe>	DanC: was it just a typo; or is it that there is no implem experience that supports it use
10:23:02 *	dajobe-jang  raises his hand
10:23:27 <scribe>	RESOLVE: lack of impl experience leads us to justify removing aboutEach*
10:24:07 <DanC>	for the record, I propose: where as the WG sees insufficient implementation experience of the aboutEachPrefix feature in RDF, it does not belong in a W3C Recommendation.
10:24:12 <AaronSw>	interesting ideas -- use aboutEachPrefix to define a container and make assertions about the container
10:24:34 <scribe>	DaveB: we were talking about aboutEachPrefix !
10:24:58 <scribe>	Brian: all agree with DanC?
10:25:04 <scribe>	All: YES!
10:25:32 <scribe>	========= AOB ========
10:25:36 <scribe>	Test case format:
10:25:42 *	danbri  agrees pathetically from irc-land
10:26:02 <DanC>	porposal http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/
10:26:17 <DanC>	rather http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001May/0264.html
10:26:44 <scribe>	DanC: Jan proposed something; I proposed something in 264
10:27:12 <scribe>	DanC: there are 2-3 batches of test cases that follow my format
10:27:17 <scribe>	Jan: I'm OK with that
10:27:35 <scribe>	Jan: while it looks like N3; please keep test cases simple
10:27:49 <scribe>	DanC: agree - 1 triple per line
10:27:49 *	GK  raise hand for question about test cases
10:28:02 <scribe>	DanC: absolute simplest form
10:28:15 <dajobe-jang>	can we add a # comment lines please
10:28:17 <scribe>	Brian: don't call it N3; it's our format
10:28:30 <scribe>	Jos: don't understand the name
10:28:36 <scribe>	DanC: could be any name
10:28:57 <DanC>	n2.5
10:29:13 <scribe>	Brian: let's take the name offline
10:29:15 <DanC>	jan
10:29:19 <dajobe-jang>	jan groans
10:29:20 <DanC>	wilber. petey. rdf-result
10:29:53 <scribe>	Brian: I move we take this off line
10:30:08 <scribe>	Brian: test cases and absolute URIs
10:30:16 *	GK  still wants to ask about test cases
10:30:25 <scribe>	... don't know the abs URI until after you sent the message
10:30:33 <scribe>	... seems like we need a test repository
10:31:41 <scribe>	Brian: EricM, will you investigate ways we can meet this need
10:32:02 <scribe>	ArtB: I don't want to be part of the pipe to get tests to the W3C's rdf-tests/ repository
10:32:13 <scribe>	ACTION: EricM: I will investigate
10:32:24 <scribe>	DanC: who wants write access?
10:32:49 <scribe>	ME: Aaron, DaveB, Brian, Jan, Martyn, ArtB, DanC
10:32:56 *	danbri  too
10:32:56 <scribe>	... DanBri
10:32:59 <em>	+ericm +danbri
10:33:05 <scribe>	... EricM
10:33:06 <GK>	What, exactly, does a test case consist of?  Is it (a) specification of some 'input', (b) spec of 'input' and corresponding 'output' (if so - how related?), (c) something else?
10:34:09 <scribe>	Jan: for the most part, we are currently dealing with syntactic issues
10:34:15 *	em  suggests rdf metadata associated with the test cases
10:34:24 <scribe>	... the output we specified should be unambigous 
10:34:32 <dajobe-jang>	metadata - I did that for my parser tests
10:34:36 <DanC>	"n3-ish"; see, we already need a name. ;-)
10:34:37 <scribe>	... it should reflect our interpretation of the spec
10:35:04 *	dajobe-jang  raises his hand
10:35:29 <scribe>	em: how about a vocab for describing the test cases?
10:35:36 <dehora>	gk's (b) seems best for automating tests: output machine readable
10:35:43 <scribe>	DaveB: I've created a schema for describing test cases
10:35:49 <GK>	Name for test case interpretation format: 'N-triples' seems to match the intended use.
10:36:05 <AaronSw>	http://ilrt.org/discovery/2001/03/parser-tests/ 
10:36:17 <AaronSw>	http://ilrt.org/discovery/swsw/pt 
10:36:22 <scribe>	ACTION: DaveB: I'll send a pointer and update it
10:36:27 *	scribe  thanks Aaron
10:36:50 <DanC>	=========== xml base
10:36:52 <scribe>	========  XML Base ==========
10:37:08 <scribe>	Jan: gives a way to attach a base URI to attributes and element content;
10:37:17 <scribe>	... doesn't say what to do with it
10:37:23 <scribe>	... two of my tests use it
10:37:40 <scribe>	... gives a way to anchor rel URIs where they occur in the serialization
10:37:52 <scribe>	DanC: would you expect existing RDF imple to support this?
10:38:05 <scribe>	Jan: not current ones; would be a useful addition
10:38:32 <scribe>	DanC: is there something in the spec that implies parsers should support this?
10:39:01 <AaronSw>	jan: this operates at the infoset level
10:39:08 <AaronSw>	danc: we need one language, one set of triples
10:39:19 <scribe>	DanC: can't define diff triples depending on whether XML base is implemented or not
10:39:57 <scribe>	DanC: my pref - this goes into the next time pile
10:40:07 <scribe>	em: who owns the next time pile
10:40:19 <scribe>	Brian: would XML base be useful?
10:40:22 <dajobe-jang>	daveb: I thought it would be a bad idea to allow now
10:40:26 <scribe>	DanC: absolutely.
10:40:50 <AaronSw>	useful, that is
10:40:51 <scribe>	... don't expect the dev community to already have supported this.
10:42:12 <dajobe-jang>	(jan pronounced with hard-"j")
10:42:16 <scribe>	DanC: it does not automatically apply everywhere [it == XML base]
10:42:41 <scribe>	ACTION: Jan: follow-up about this on the list
10:42:50 <GK>	I agree with DanC that changing the meaning of currently-valid and widely understood RDF would be the worst kind of backward-incompatibility
10:43:31 <scribe>	Jan: I will summarize what it might do; if too great of an impact, it should go on the next time list.
10:44:33 <scribe>	Jos: TBL mentioned caching, would someone explain that?
10:44:42 <DanC>	Jos, I use wwwoffle
10:44:43 <scribe>	DanC: I'll talk to you offline about that
10:45:00 <scribe>	======== aboutEach ==========
10:45:17 <scribe>	DaveB: should we get rid of aboutEach?
10:45:24 <scribe>	DanC: short answer: yes.
10:45:47 <scribe>	... I posted my interpretation; I could put it in the next time pile
10:46:03 <scribe>	DaveB: I agree with DanC; should be layered
10:46:09 <scribe>	DanC: has anyone used it?
10:46:19 <scribe>	AaronSw: I've used it a little
10:46:25 <scribe>	Brian: what is the problem?
10:46:47 <scribe>	DanC: the spec isn't clear
10:47:02 <scribe>	Brian: I think it is clear - but it's syntactic sugar.
10:47:13 <scribe>	Ora: I use it, don't want it removed;
10:47:47 <scribe>	Brian: my discussions with DanBri have been along the lines of keeping it
10:48:01 <scribe>	DaveB: if have a hugh file, can cause implementation problems
10:48:18 <scribe>	Ora: it's a linear issue
10:48:50 <scribe>	ACTION: DanC: point out the spec bug - need an errata
10:49:17 <scribe>	DaveB: aboutEach makes it difficult to implement wrt streaming
10:49:28 <scribe>	em: what does it mean to take something out and leave something in?
10:49:45 <scribe>	DanC: can't be generalized
10:50:08 <scribe>	... we decided before that lack of impl experience was justification to remove something
10:50:17 <scribe>	Who is using it and how?
10:50:26 <dajobe-jang>	(said daveb)
10:50:27 <scribe>	Jan: I use it but it is problematic
10:50:42 <scribe>	Ora: I use it; wasn't difficult; 
10:51:04 <scribe>	ACTION: Ora: post imple details for aboutEach
10:52:12 <scribe>	DanC: does your impl work if you use _<n>?
10:52:15 <scribe>	Ora: yes
10:53:07 <scribe>	Brian: DaveB, are you OK?
10:53:19 <scribe>	DaveB: I want to read Ora's report.
10:53:35 <scribe>	Brian: anything else?
10:53:48 <scribe>	======= Anything else? ========
10:54:08 <scribe>	DanC: do you folks use the W3C meeting registration form?
10:54:43 <scribe>	... it can be useful
10:55:14 <scribe>	DanC: what are the expectations for the f2f?
10:55:27 <scribe>	Ora: informal social interaction is very useful.
10:55:35 <dajobe-jang>	daveb: TIN - Tin Isn't N3
10:55:38 <DanC>	rdf-result
10:55:41 <DanC>	ntriples
10:55:43 <scribe>	=========== Test format name ==========
10:55:53 <DanC>	n-triples, says GK
10:56:01 <dehora>	linear form but TIN is a good name  :)
10:56:13 <scribe>	ArtB: I vote for TIN
10:56:25 <em>	+1 tim
10:56:28 <scribe>	This Isnt N3 == TIN
10:56:30 <em>	s/tim/tin
10:56:32 <DanC>	I like n-triples
10:56:36 <dehora>	+1 TIN
10:56:51 <scribe>	Jos: I like n-triples too
10:57:08 <scribe>	Brian: motion to use n-triples
10:57:12 <scribe>	DanC: seconded
10:57:13 <dajobe-jang>	+2 jan, daveb
10:57:28 <scribe>	RESOLUTION: n-triples it is!
10:57:42 <scribe>	Next meeting: Jun 08
10:58:07 <scribe>	Scribe next week: Jan Gran
10:58:15 <scribe>	Adjourned ...

Received on Friday, 1 June 2001 12:30:46 UTC