RDF Core WG Minutes 11 May 2001

This page summarizes the telecon for the RDF Core Working Group held from 1400 to 1500 UTC on 11 May 2001. These minutes were prepared by Dave Beckett, but have not yet been reviewed by the other participants.

Meeting IRC Log


Regrets received from Dan Brickley, Mike Dean, Bill de Hora.


Brian McBride added issue from XML Base discussion on www-rdf-interest this week to the end of the agenda

Review Minutes of Previous Meeting

The previous minutes of 27 April 2001 were accepted.

Review of Actions

Issues Progress

The RDF Model and Syntax spec is unclear about when rdf: prefix is needed

Issue #rdf-ns-prefix-confusion

Dave Beckett asked Art Barstow if he was happier with the revised wording Dave proposed and Art agreed that the list of results were acceptable.

Art asked about the second issue Dave introduced and after discussion it was agreed that the proposed answer 1) was the best alternative, especially since the current grammar forbids unprefixed attributes. Jan grant noted M+S grammar productions 6.6, 6.7, 6.8 need updating amongst others and Dave was asked to add test cases of how the proposed change would work.

ACTION Dave Beckett: Update the proposed changes for this issue and add test cases to demonstrate how they worked.


Issues #rdf-container-syntax-ambiguity and #rdf-containers-syntax-vs-schema

Dan Brickley had sent a message to www-archive, that not everyone had seen. Ora said he would like to revise it and post to rdfcore list when ready and would like feedback then, so that the issue can be discussed.

Ora noted that it was intended that containers would be considered to be moved out of the core data model.

ACTION Ora Lassilla/Dan Brickley: send analysis to rdfcore-wg list

ACTION All: Respond to Ora/Dan's notes when they appear.

General editorial comments.

Issue #rdfms-editorial

ACTION Ora Lassilla: Carried over (Dan Brickley and Guha previously recorded to help)

The interpretation of empty property elements is unclear

Issue #rdfms-empty-property-elements

Jan Grant summarised his analysis but basically the answer is, it is an empty literal, and although some people suggested an anonymous resource, that is a bit broken. Consensus on this.

RESOLVED: The answer is an empty literal (wordsmithing needed on M+S text)

ACTION Jan Grant: reword and send a new message with test cases, including one with just an ID attribute.

What mime type should RDF Schema and other RDF documents have?

Issue #mime-types-for-rdf-docs

Aaron Swartz outlined the document got a favourable response apart from the interaction with fragments and assertions. Not everyone was still happy with this and since it made statements about what RDF meant, maybe not good in a mime types doc? Brian asked if there was urgency for this since it wasn't clear if this could be published and amended easily (unsure of IETF mime type process). Agreed to postponed it until further work done on fixing various things that it depends on.

Not ID and resource attributes

Issue #rdfms-not-id-and-resource-attr

Dave Beckett note that the existing analysis had a flaw and would need revising, would do for next meeting.

ACTION Dave Beckett: Revise analysis and post to list

MUST a parser created bags of reified statements for all Description elements?

Issue #rdfms-reification-required

Skipped, Guha not present. Later on Guha arrived and said he will have something before the next meeting.

ACTION Guha: Present analaysis to list for discussion.

What is a resource and how does it relate to other concepts such as URI and entity?

Issue #rdfms-resource-semantics

Long discussion after Martyn Horner's posting to the list (just before meeting) defining three terms - URI, Entity and Resource. Martyn presented the ideas and after discussion, request for more formalisms from Jos De Roo, Martyn a good goal but a part of the things we want to do is expressing ambiguity. Frank [which one?] said formalisms are good but need to be expressed from strong ideas.

Discussion of binding resources and URIs, various views were expressed that were not consistent. Ora noted that original RDF M+S people punted this issue. Some discussion of Resource=URI but need to have more discussion on-list. Dan Connolly emphasised the need for using test cases to base the discussion - see also his comments to the list on 8 May and 11 May.

Digression into test cases and Brian, Jos, Dave resolved strongly to use these. Jos wondered about data for testing and correct results of, say, parsing but Dave responded that there was at present no canonical triple dump syntax so this wasn't yet possible.

RESOLVED: Make all test cases / results public on the web.

ACTION Brian McBride: Link test cases, results etc. from issues list

The RDF community needs a precise terminology to enable it to discuss issues.

Issue #rdf-terminologicus

Discussed in previous issue.

XML Base

Brian outlined XML Base is now a candidate recommendation and there had been discussion this week on www-rdf-interest about this including replies by Aaron and Dave. He asked does this have an impact on RDF? A discussion followed which seemed to indicate it would break old parsers if used in new RDF/XML. Brian said his working model on backward compatibility was that what we end up with might not go through existing parsers and it was better that existing parsers failed on new syntax. Dave disagreed and wanted the existing syntax to continue to work. Jan Grant offered to do an analysis and summarise it to the list.

ACTION Jan Grant: Do an analysis of the impact of XML Base and summarise to list.

Next Teleconference

There will be a teleconference at the same time next week - 1400 to 1500 UTC on 18 May 2001.