W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2001

Minutes: Syntax subgroup teleconf - Oct 26, 2001

From: Art Barstow <barstow@w3.org>
Date: Fri, 26 Oct 2001 10:17:45 -0400
To: w3c-rdfcore-wg@w3.org
Message-ID: <20011026101744.A16212@w3.org>
What: 	RFD Core WG Syntax teleconf
When: 	Oct 26, 2001
Who: 	Jeremy, DaveB, Brian, Steve, Eric, Art

The dicussion focused on whether the RDF/XML Syntax doc - editor's
version is at:

 [1] http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/ 

should define the syntax to triple transfomration in "Structured
English" [as is currently done in [1]] or to use the method Jeremy
has described in:

 [2] http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/0326.html


  That we will use structured English [in the Syntax document] to 
  describe the transformation from the grammar to n-triples.  The 
  rule based formalism [done by Jeremy] will be used to verify 
  that the structured English is correct.

IRC log:

Session Start: Fri Oct 26 08:32:11 2001
[08:32] *** Now talking in #rdfcore
[08:32] *** Topic is 'RDF Core WG Meetings 13:00 - 15:00 UTC Friday[08:32] *** Set by dajobe on Fri Oct 26 08:39:54
[08:40] *** jjc has joined #rdfcore
[08:40] <jjc> hi art
[08:44] *** dajobe has joined #rdfcore
[08:45] <ArtB> 'morning gents
[08:45] <jjc> morning
[08:45] <ArtB> Dave - I'm just starting to look at syntax ver 1.95.  Is that the primary topic for this meeting?
[08:45] <dajobe> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/0504.html
[08:46] *** spetschu has joined #rdfcore
[08:46] <spetschu> hi folks -- is anyone on the syntax call?
[08:46] * ArtB is not (yet)
[08:46] * dajobe dialed
[08:46] <jjc> just joined
[08:47] <dajobe> yes, above is one of the things
[08:47] <jjc> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/att-0408/01-tp-jjc-no-xslt.html
[08:47] *** danb-lap has joined #rdfcore
[08:49] <dajobe> artb dialling?
[08:50] <ArtB> +ArtB
[08:51] <spetschu> +SteveP
[08:51] <dajobe> bwm can't scribe just yet - rebooting
[08:52] <spetschu> Item 1 - Jeremy describe his doc, group discuss
[08:52] <ArtB> jjc: the snail work will tell where triples come from, not just give the triples
[08:52] <dajobe> Jeremy is trying to provide a formal description of where triples come from
[08:53] <dajobe> art: I'd be happy if you'd scribe
[08:53] <ArtB> ok
[08:54] * ArtB is having trouble hearing jjc - his speech is broken ...
[08:55] <ArtB> dave: I'm concerned it doesnt meet our target audience (it = the snail doc)
[08:55] <ArtB> steve: is a goal to have consistency b/w the 2 docs?
[08:56] <ArtB> jjc: we need to describe where the tripels come from
[08:56] <ArtB> jjc: choices: 1. English text (ala M&S); 2 a sep set of rulles 
[08:56] <ArtB> ... rules describe what triples are generated
[08:57] <ArtB> daveb: the separation may be too difficult for people to deal with
[08:57] <ArtB> daveb: I don't wat to see the separation
[08:58] <ArtB> daveb: XPath can be used as a template for keeping things precise
[08:58] <ArtB> artb: I haven't looked at jjc's document
[08:59] <ArtB> steve: I scanned jjc's doc; read dave's
[09:00] <ArtB> Brian: we need to decide on an approch but more WG memebers need to read JJC's doc
[09:01] * ArtB can't hear whoever is speaking
[09:02] *** em has joined #rdfcore
[09:02] <ArtB> Brian: I think the doc isn't too big
[09:02] <ArtB> Brian: the intro text is a little off-putting
[09:02] <ArtB> ... but that can be fixed
[09:03] <ArtB> ... I find it hard to know if we can describe alal of the cases in English
[09:04] <ArtB> ... seems like it would be easier to get all of the case by describing a transformation
[09:04] <ArtB> Steve: the audience?
[09:04] <ArtB> dave: we decided its parser writers
[09:05] <ArtB> Steve: want to hear about Dave's doc so we can make a better decision
[09:05] <ArtB> BM: yep, sounds good
[09:06] <spetschu> http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/#section-Updated-Grammar-changes
[09:06] <spetschu> http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/
[09:06] <jjc> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/att-0393/01-index.html
[09:07] <ArtB> Dave: shows a good link from XML and go into model used in XPath
[09:08] <ArtB> Dave: approach - for each production, would write a description of the triples that would be generrated
[09:08] <ArtB> ... I wrote down some of the triples that would be generated
[09:08] <ArtB> ... gets around some of the bnode naming problems
[09:09] <ArtB> ... the doc has changed some
[09:09] <ArtB> ... the goal was for the 2nd WD to have the grammar specified and to start the mapping
[09:10] <ArtB> ... the 3rd WD would finish the mapaping
[09:10] <ArtB> ... the 4th would be cleanup
[09:10] <ArtB> Steve: reader must understan XPath, Infoset, RDF, XML, ...
[09:11] <ArtB> Dave: don't need to understand XPath but do need to know SAX
[09:11] <ArtB> JJC: reader must know the XPath node set
[09:13] <ArtB> ArtB: it would be good to explicitly state what background docs the reader must understand to grok the syntax doc
[09:13] <ArtB> BM: does anyone want to urge the editor to change his preference
[09:14] <ArtB> scribe: no objections
[09:14] <ArtB> EM: spit the question - have they read enough to make a deciosoon; if yes, what do you recommend
[09:15] <ArtB> Steve: who has readd enough of both docs to make a decision?
[09:16] <ArtB> Steve: I can make a high-level decision
[09:16] <ArtB> EM: I haven't read enough to make a recommendation
[09:16] <ArtB> ... I echo the point about how muchwe agree on what triples should be emmitted
[09:20] <ArtB> BM: I don't want to see us delay any longer
[09:20] <ArtB> ... we have to proceed 
[09:21] *** bwm has joined #rdfcore
[09:21] <ArtB> JJC: I think we shoujld have 1 or the other, not both
[09:22] <ArtB> EM: yes, but we need to make things as comprehensive as possible
[09:22] <ArtB> ... Could JJC's stuff feed DB's document?
[09:23] <ArtB> ... that is, JJC continues his work and make sure DB's document covers all of the cases
[09:23] <bwm> PROPOSE: That we will use structured English to describe the transformation from the grammar to n-triples.  The rule based formalism will be used to verify that the structured english is correct.
[09:24] <ArtB> DB: I would be happy with that
[09:24] <ArtB> +1 from ArtB
[09:25] <ArtB> JCC: I would be somewhat disappointed on an emotional level
[09:25] <ArtB> ... the lack of response so far reflects that perhas thw WG doesn't think my work was as important that it should be
[09:27] * ArtB wonders if it would be useful to get feedback from "real" parser writers (e.g. Meggison, CARA, Jason Diamond, etc)
[09:27] * em agrees with ArtB with this
[09:28] * em will also note however that we have in part a RELAX schema based on the first draft approach
[09:29] * em raises his hand
[09:32] <ArtB> EM: if someone was starting from scratch, I wonder which approach would be more useful
[09:33] <ArtB> Steve: does everyone agree with BM's PROPOSE above?
[09:33] * em agrees with bwm's proposal
[09:33] <ArtB> DB: OK with me
[09:33] <ArtB> EM: OK
[09:33] <ArtB> AB: OK
[09:33] <ArtB> JJC: Abstain
[09:34] <ArtB> Steve: any objections?
[09:34] <ArtB> scribe: no objections raised
[09:34] <spetschu> closed!
[09:40] *** ChanServ sets mode: +o dajobe
[09:41] *** Disconnected
Session Close: Fri Oct 26 09:41:10 2001
Received on Friday, 26 October 2001 10:17:44 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:05 UTC