· Logical Data Description and Format Description have been placed in review as received from the Data Description working group without additional review by the Modeling Team.
· The Methodology Pattern has been placed in review without a specific instantiated application. Methodology Pattern will be found only in the model, not in the XML Bindings as its classes are all abstract. Several instantiations of this pattern are anticipated in the upcoming special Codebook Release.
· The interaction of Patterns (Collection, Process, Methodology, and Signification) is still under discussion. In particular, questions about combining Methodology and Process Patterns and the detail of content within classes belonging to a pattern are still being discussed. Comments and recommendations are welcome. As this is resolved the Modeling Team will provide a high level document providing a bird’s eye view of the conceptual framework of DDI, how the patterns interact, and how the patterns work together to express similar relationship patterns within different areas and applications of DDI.
· Data Capture and Instrument View provide a design-time example of the Process pattern. We are working on a run-time example within the broader needs of a run time layer that provides actual run time binding, plus action logs, variation from design etc. This is part of the temporal perspectives talked about before (looking forward/design, execution/run, looking backward/what took place). This plays into the issue of capturing processing provenance as well as paradata during data capture operations.
· The Service class in the Process Pattern will not talk about time. This will appear in the concrete class that realizes Service. There are three applications of time: Effective time (Availability), Historic time, and Expected Time. Look at estimatedDuration in the WorkflowService object.
· Service does not relate correctly to Process Step. Leave as an open issue. Service needs a more complete review, but we should wait until we see how people implement it (regarding what properties they need.) There are also issues with the relationship between process step and service including the directionality of the relationship and the use of internal controlled vocabularies.
· Should the Variable Cascade approach be replicated for forms of capture such as Questions? Is there a parallel in terms of comparison and reuse? If InstanceQuestion inherits from RepresentedQuestion it would solve the current issue of requiring the creation of both a RepresentedQuestion and an InstanceQuestion.
· The Variable Cascade has been modified since the last development release and should be reviewed carefully.
· Only limited physical formats are currently described (Fixed Format and CSV). Development of other formats may include additional properties for PhysicalLayout, ValueMapping, and ValueAndConceptDescription.
· Datum and its consequences are still being worked out. It expands the scope of DDI from the record into the actual datum (observation) and allows for support of datum specific paradata and viewpoints.