RDF Core WG Syntax subgroup minutes for telecon 2001-10-05

RDF Core WG Syntax subgroup minutes for telecon 2001-10-05

Transcript:
  http://ilrt.org/discovery/chatlogs/rdfcore/2001-10-05


Present:
  Art Barstow
  Dave Beckett (chair, co-scribe, minutes)
  Dan Connolly
  Jeremy Carroll
  Jan Grant (co-scribe)
  Pat Hayes (joined later)


0: Allocate scribe:  Dave (and Jan later)

1: Review Agenda (typed into IRC)

  No AOB

2: Schema validation technology - what to use to validate RDF/XML?

A strawpoll saw general approval that the non-XML RelaxNG schema
produced by James Clarke was popular:
  http://lists.w3.org/Archives/Public/www-rdf-comments/2001JulSep/0238.html

Discussion of whether having a normative validation schema in the
specification was a good idea.  Agreement that prose would describe
what is valid declaratively and the existing WD is reasonably
intelligible.  Agreement that the target audience of the syntax
document is parser writers, not end users - we require something they
can review and understand.  The primer must cover the rest of the
explanation of the syntax.  A potential approach with general
agreement is to have an accurate description of syntax in prose with
a non-normative RelaxNG and other schemas in appendixes.


3: Transformational approaches from RDF/XML to N-Triple

Jeremy outlined his approach: transforming RDF/XML using XSLT and
staying within XML for as long as possible because that's the natural
setting for thinking about the abbreviated syntax as abbreviations.
The XSLT he writes is generated from a smaller rules language, that
apply XML transforms anchored to the root XML element (rdf:RDF) to
flatten the structure This avoids syntax striping issues (since nodes
are always below root).  bNodes have to be introduced and done using
beyond-RDF/XML syntax to distinguish from nodes with URIs.  Jeremy
explains that the language is not very stable at this point and
estimates it will have 2x rules in a complete version (reification).

ACTION 2001-10-05#1 Jeremy Carroll:
   Email work to w3c-rdfcore-wg list as attachments to get archived.

Discussion of rdf:aboutEach and how it is still broken, not really in
syntax or model.  Dan refers to his rdf:aboutEach test case
conundrum:
http://lists.w3.org/Archives/Public/www-rdf-comments/2001AprJun/0107.html

Discussion of the XSLT approach; Jeremy wants not to specify the
mapping in a "programming language", Dan thinks XSLT is one since it
is Turing complete.  Jan says yes, but what matters is that the
mapping is clear, whatever language.  Jeremy explains the rules.xml
is intended to be compiled into HTML info "visually appealing"
patterns.  It it the generated HTML that is meant to make the rules
clear, and would be normative, indicating that applying the
transformation rules in any order gives the same result.

Dan explained he wrote an RDF/XML parser in XSLT in April (in an
afternoon):
  http://www.w3.org/2001/04rs22/
Although not complete, and it generated a lot of issues on the issues
list.  Would be similar using another approach such as attribute
grammar (Brian McBride did one) but although kind of long to do in
XSLT, once familiar, the idioms are rather straightforward.  Dan
notes there is XSLT support for ID generation for introducing
identifiers for bNodes.

Discussion of future work.  Dave notes he'll continue with existing
style but try going for a more sax-like Infoset trying to make an
EBNF with the infoset terms as terminals.   Jeremy notes he has
an XML version of this that could be useful to generate this.

ACTION 2001-10-05#2 Jeremy Carroll:
   Email XML version to Dave and/or the w3c-rdfcore-wg mailing list.


The group notes that the HTML will be normative and any non-normative
XML appendix of mapping etc. is a pleasant bonus.

Agreement to plough on with existing approaches (Dave, Jeremy)


4: Outstanding issues from issues list

Dave outlines the goal of looking at the issues on the issues list at
  http://www.w3.org/2000/03/rdf-tracking/
and move the job of writing their resolution to dealt with by the
document editor(s) and workers, to speed up work and do it attached
to working on documents.  The issue list would point to resolutions
still recorded in the issues list.  The WDs would have a section
listing the issues resolved, addressed, discussed by the document,
along with pointers to the words that address them.

Then the group did a walk through the Issue Tracking doc
  http://www.w3.org/2000/03/rdf-tracking/
"Currently Active Issues" section, sub section "Model and Syntax Issues"
and made the following split between syntax, non-syntax issues:

  4.1 Purely Syntax Issues

  Issues with no resolution yet:

  http://www.w3.org/2000/03/rdf-tracking/#rdfms-aboutEach-on-object
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-syntax-desc-clarity
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-quoting
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-nested-bagIDs
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-qnames-cant-represent-all-uris
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-replace-value
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-propElt-id-with-dr
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-formal-grammar
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-xml-literal-namespaces
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-qname-uri-mapping
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-qnames-as-attrib-values
   -- (aside: answer is "No")
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-validating-embedded-rdf
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-xml-base
  http://www.w3.org/2000/03/rdf-tracking/#mime-types-for-rdf-docs
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-abouteach
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-reification-required
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-rdf-names-use

  Active Issues:
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-not-id-and-resource-attr
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-difference-between-ID-and-about

  Resolved Issues:
  http://www.w3.org/2000/03/rdf-tracking/#rdf-ns-prefix-confusion
  http://www.w3.org/2000/03/rdf-tracking/#rdf-containers-syntax-ambiguity
  http://www.w3.org/2000/03/rdf-tracking/#rdf-containers-syntax-vs-schema
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-abouteachprefix
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-empty-property-elements


  4.2 Partial Syntax Issues

  http://www.w3.org/2000/03/rdf-tracking/#rdfms-editorial

  These are general editorial comments on RDF M&S, some of which will
  apply to new syntax documents.


  4.3 Non Syntax Issues in "Model and Syntax Issues" section

  http://www.w3.org/2000/03/rdf-tracking/#rdfms-resource-semantics
  http://www.w3.org/2000/03/rdf-tracking/#rdf-equivalent-uris
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-fragments
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-literalsubjects
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-xmllang
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-contexts
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-identity-of-statements
  http://www.w3.org/2000/03/rdf-tracking/#rdf-containers-otherapproaches
  http://www.w3.org/2000/03/rdf-tracking/#rdf-equivalent-representations
  http://www.w3.org/2000/03/rdf-tracking/#rdf-formal-semantics
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-seq-representation
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-assertion
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-logical-formalism
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-logical-terminololgy
  http://www.w3.org/2000/03/rdf-tracking/#rdfms-boolean-valued-properties

  Note: We didn't decide whether they were Model issues, just that they
  were non-Syntax.

  4.4 For all issues, review decisions as how they affect the syntax.

  Examples include rdfms-xmllang, anything to do with literals or the model.


ACTION 2001-10-05#3 Dave Beckett: 
   Looking at potential resolutions for all the undecided syntax
   issues above.


Then the group had a discussion of where the resolution goes along
with words in syntax docs.  The resolution would be recorded in
issues list, with words in the syntax docs too; test cases for the
issue resolution go in the usual test cases area and document, linked
from the syntax doc and issues list.  Dave noted that the curent
refactoring syntax document lists several issues and how they caused
changes in

  http://www.w3.org/TR/2001/WD-rdf-syntax-grammar-20010906/#section-Updated-Grammar-changes

The intention is that new versions of this document would say "we're
address, explain, discuss, ... this issue in section X"

Discussion of when to get next WD out (RSN of course).  agreement it
would be good to have most of disposition done by Christmas 2001
rather than aiming for WWW2002.

END OF MEETING

Received on Tuesday, 9 October 2001 09:44:18 UTC