Monday, October 17, 2016
Morning
Session 1 (plenary): Introductions, goals of the workshop, and overall introduction to DDI
Introduction of participants
Introduction to DDI
- DDI-Codebook
- DDI-Lifecycle
- DDI Moving Forward Project
- What do we work with and the challenges that we’ve encountered, especially regarding interoperability with other specifications
Session 2 (plenary): Presentations from external experts on specific specifications or usage of specifications.
External presentations:
- Klas Blomqvist – GSIM
- Flavio Rizzolo – LIM
- Arofan Gregory – SDMX
- Eric Prud’hommeaux – FHIR
- Barrie Nelson - CDISC
Afternoon
Session 3 (plenary): Presentations from external experts on specific specifications or usage of specifications.
External presentations:
- Nick Car – PROV
- Gregg Kellogg – CSVW
- Gary Berg-Cross – RDA & Ontologies
- Daniella Meeker – standards in biomedical research
- Darren Bell – UKDA’s use of DDI in Big Data
Session 4 (plenary): Discussion of interoperability
Interoperability
- Presentation by Wendy Thomas
- Begin discussion towards an agreed definition of interoperability
Evening:
Possible evening session - Topic to be determined
Tuesday, October 18, 2016
Morning
Session 1 (plenary): Discussion of interoperability
- Ensure we have an agreed upon definition of Interoperability
- Begin discussion of business scenarios
- Short presentations of requested business scenarios (Jay, Eric, Jon, Barry)
- Collection of business scenarios
Session 2 (working groups):
Session 2A: Business scenario: Situations where we would use different standards together
- Outcome: Document outlining these business scenarios
- Description of the Business Scenario in a consistent way
- Agree on a set of functionalities across the specifications
- List of requirements
Session 2B: Common Aspects of all specifications present during Data Description
- Outcome: 2 tables:
- High level table demonstrating what specs are used for specific aspects of … see Arofan’s table
- Detailed table mapping the important properties and relationships within classes, terminology differences, identify significant differences
Afternoon
Session 3 (working groups):
Session 3A: Provenance of data, metadata
- Common understanding of what provenance is
- What are the requirements of provenance
- Overview of real life provenance applications
Outcome: A common definition / understanding of provenance for the purpose of this workshop
- Discussion of provenance
- Real life examples
- Shared and agreed definition of provenance
Session 3B: Design patterns
- Review of practical use in DDI 4
- Generic principles, what is the right balance between clean structure on the model level and understandable use on the binding level
- Pros/cons on model level, binding level
Outcome: document outlining:
- Evaluation of current design patterns in DDI4
- What can we learn from the other specifications
- Reason for pattern use and style of pattern
Session 4 (working groups):
Session 3A continued: Provenance of data, metadata
Session 3B continued: Design patterns
Evening:
Possible evening session - Topic to be determined
Wednesday, October 19, 2016
Morning
Session 1 (working groups):
Session 1A: Bindings of DDI, CDISC, HL7/FHIR, SDMX, GSIM, CSV on the Web
- Approaches of modeling and/or generation
- Usage of other vocabularies
- The process you use to create your specifications
Outcome: document outlining why other specs have chosen their particular bindings, the process used to create them, their rationale, and approaches to new releases and versioning; how to deal with issues of identification and round-tripping of data and metadata
Session 1B: Operations (Protocols) with specifications, i.e. REST with DDI, CDISC, HL7/FHIR, SDMX, GSIM, CSV on the Web (Part 1)
- Goal is to find a generic way of operations to support combined usage of specifications
- Requirements from a user’s perspective
Outcome: document contrasting the approaches used to interchange protocols (REST, SPARQL, SOAP, etc..) by the different specifications; Proc and cons of having standard interfaces as part of the specification and the rationale
Session 2 (working groups):
Session 1A continued: Bindings of DDI, CDISC, HL7/FHIR, SDMX, GSIM, CSV on the Web
Session 1B continued: Operations (Protocols) with specifications, i.e. REST with DDI, CDISC, HL7/FHIR, SDMX, GSIM, CSV on the Web
Session 2 (plenary):
Discussion – to determine if there is a consensus view of which bindings/protocols should be used; what does this picture look like? (Current and future perspective - possible developments)
Afternoon
Session 3 (working groups):
Session 3A: Provenance of data/metadata (Data production)
- Relationship of stages of the data life cycle
- Relationship of tasks in GSBPM/GLBPM, other process models
- Relationship of aggregate data to underlying microdata – description of processing
How to record provenance; Can PROV-O describe other process models such as GSBPM/GLBPM? How can the ideas of PROV-O be used in non-RDF bindings?
Outcome:
- Quick presentation to the larger group on discussion and findings
- Document from presentation to be used in joint doc
Session 3B: Provenance of data/metadata – RDF (Data publication)
- Usage of W3C PROV-O in DDI 4 RDF and other specs with RDF bindings
- How do we capture sufficient information around provenance and processing?
Goal: best practice for using PROV-O with other Specs
Outcome:
- Quick presentation to the larger group on discussion and findings
- Document from presentation to be used in joint doc
Session 4 (plenary):
Plenary session for groups to report back to exchange findings and open discussion
Outcome: Extension of provenance definition document to include:
- Data production discussion
- Data publication discussion
Evening:
Possible evening session - Topic to be determined
Thursday, October 20, 2016
Morning
Session 1 (working groups):
Session 1A: RDF bindings for statistical data and metadata
- Discuss requirements for RDF applications that deal specifically with statistical data
- What are the best RDF vocabularies for data and metadata description?
- Specific best practice for using multiple vocabularies for this purpose
Outcome: document describing the requirements and best practices for use of RDF vocabularies in statistical applications
Session 1B: Data description, general mapping table of multiple specifications with equal and similar items
- Commonalities across DDI, CDISC, HL7/FHIR, SDMX, GSIM, CSVW - data description aspect
- More detailed discussion as follow-up to Tuesday’s discussion
- Opportunity for folks not involved on Tuesday to participate
Outcome: document outlining the relationship of different standards for data description
Session 2 (working groups):
Session 1A continued: RDF bindings for statistical data and metadata
Session 1B continued: Data description, general mapping table of multiple specifications with equal and similar items
Afternoon
3rd Session (plenary):
Session 3A: DD-R package
- discuss and document program for moving forward on a Data Description package in R
Session 3B: RESTful
- Discuss and document information for moving forward on a collaborative RESTful
4th Session (working groups):
Session 4A: RDF Bindings and Production framework with Elisa Kendall - via remote
Session 3A continued: DD-R Package
Session 3B continued: RESTful
Friday, October 21, 2016
Morning
1st Session (working groups):
Session 1A: Start the conversation for future collaboration – create a small working group to work out the details
- Multiple specs use case – work in the future?
- Think about active proposals for future collaboration
Session 1B: Statement of design principles to be agreed by specs represented (Data Manifesto)
- How do we make available?
- How/where do we publish?
Outcome:
The outline and draft of agreed design principles (“Data Manifesto”)
2nd Session:
Afternoon
3rd Session (plenary):
- Communication and sustainability