Observations Workshop 2007 Session 3

This is version 5. It is not the current version, and thus it cannot be edited.
[Back to current version]   [Restore this version]

Session 3: Requirements for a shared observational data model

Breakout-Group Products

• A list of requirements for a shared observation data model
• Include modeling features, enabling capabilities, and architectural and adoption constraints

Background

Based on the overall merit and feasibility of developing a shared model for observational data, this session aims at enumerating the specific requirements for the creation and use of a shared model of observational data. Session topics should focus on the: (1) the key characteristics of an observational data model; (2) architectural issues of the data model; and (3) representation technologies for the data model.

Discussion topics

Which aspects of observational data and associated capabilities are best captured and supported by a shared observational data model? Which are better captured and supported within domain-specific extensions of the shared model?

Potential issues:

• What are the key characteristics of a shared observational data model?
• What high-priority features should comprise a minimal, base shared model?
• What lower-priority features should be included (for later inclusion)?
• Should the model inherently support extensibility, and to what level?
• What base capabilities should be supported? What constraints do these place on the data model?
• What are the primary architectural requirements?
• What modeling language should be used to define the shared data model, and why? (e.g., description logic/Semantic Web languages RDFS, OWL-Lite, OWL-DL, OWL-Full? XML schema? UML? multiple languages?)
• What formats should be supported (for expressing information in the model)?
• What other types of constraints should be addressed (e.g., incentives for adoption, access rights, etc)?

Session notes and results