- From: Brian McBride <bwm@hplb.hpl.hp.com>
- Date: Fri, 03 Aug 2001 14:49:16 +0100
- To: rdf core <w3c-rdfcore-wg@w3.org>
For the record, this is a transcript of the various flip charts written at the face to face. Brian ------------------------------------------------ Problems [rss] o Spec -> screams o namespace prefixes o need production quality RDF tools o need provenance - this chunk of xml came from ... o need unsquish o hackers don't read the spec Prism o users of RDF are meta data designers o hard to extend o reification is hard ------------------------------------------------ DAML o domain/range constraints o subclass should be cyclic o tagging statements and clumps o closed collections o formal semantics o what do uri's mean o xml schema data types ZAP o don't want to support reification o stronger datatyping o info in schema for UI support ----------------------------------------------------- DC o trying to be both XML and RDF o language support xml:lang o relationship to XML schema o ugly syntax o containers - modelling guidelines o datatyping - numbers and dates o NEED SOMETHING SOON o AI trap o No reification ------------------------------------------- N-triple needs to represent the doc -------------------------------------------------- [Issues with Abstract syntax and model] o parse same doc twice - are anon nodes the same? o split doc -> are anon nodes the same? o alternatives interpretations of reification o syntactic representation of doc ++ o formal semantics o anon resources + o literal is XML structure ++ o graph o uri-substructure ----------------------------------------------------------------- [about Prince nodes] They are not needed for assertion hard to create unique ids lazy information loss [provenance] ---------------------------------------------------- [why do we need anon resources] Names -> matters Elegance Referreral Query Don't know URI pool scheduling -> practical implementation avoid philosophical rat hole ----------------------------------------------------- <rdf:Description> <foo:bar>foobar</foo:bar> </rdf:Description> .... <foo:bar> "foobar" . --------------------------------------------------------- Is there a way to point to the distinguished node? P(x) -> exists y P(y) ---------------------------------------------------------- exists(bob) (Old(bob)) ------------------------------------------------------- Ntriple is central - model RDF/XML, model theory and graph defined in terms of it. --------------------------------------------------------- Graph is central, n-triple, model theory and RDF/XML defined in terms of it [this is the one we agreed to] --------------------------------------------------------- Schema - 6th Sept Syntax Model ---------------------------------------------------------- Two RDF documents are equivalent when they map to the same graph. ------------------------------------------------------------ Two RDF Graphs are the same when: o they are graph isomorphic o no two nodes in the graph have the same URI ----------------------------------------------------------- Literal = string parseType string or Qname namespaces language base uri entities ---------------------------------------------------- Options Stand Alone XML -> limitations parser adds namespaces -> incomplete infoset as RDF serialized infoset fragments deprecate and use cdata ------------------------------------------------------ Model Theory is a Tool which is not useful for everyone Suggestion that it can be translated to English Document Structure Consider Audience Don't overload the term spec ------------------------------ Document Structure (Hypertext?) Primer 5+2 Model 8 RDF/XML 4 Schema 4 Test Cases 2
Received on Saturday, 4 August 2001 10:40:32 UTC