- From: Eric Neumann <eneumann@teranode.com>
- Date: Tue, 17 Oct 2006 19:48:58 -0400
- To: Bo.H.Andersson@astrazeneca.com, Kerstin.L.Forsberg@astrazeneca.com, "James McGurk" <james_mcgurk@merck.com>, "Ted Slater" <ted.slater@pfizer.com>, "Uwe Trinks" <Uwe.Trinks@sentrx.com>
- cc: public-hcls-dse@w3.org
Some useful patterns to keep in mind for our next TC, when working with the SDTM and JANUS data models: - N-Ary relations (can be blank nodes) can be used to establish relations between 2 or more sets of data: complex data sets - Use of subclasses and subsumption to model data types, rather than tables : e.g., <Intervention stop_trials> a <owl:class Intervention_type> - Require tables (with RDF mappings) for only large sets of things: events, findings, interventions, studies, subjects, etc; things like clinic sites, arms, decision stages can be done via small RDF graphs that reference table elements via a URI - Relations between an event and a finding can be done using standard RDF predicate statements: no new cross table needed! - Non-repudability and authenticity : digitally sign findings with evaluator (password) and tag as a private/public keyed hash value I also want you to inform everyone that the new HCLS-DSE mailing list <public-hcls-dse@w3.org> is now active, and you'll need to subscribe to it: <public-hcls-dse- request@w3.org> Subject: subscribe Eric
Received on Tuesday, 17 October 2006 23:49:40 UTC