Observations Workshop 2007 Session 2

This is version 10. It is not the current version, and thus it cannot be edited.
[Back to current version]   [Restore this version]

Session 2: Enabling capabilities of a shared observational model

Breakout-Group Products

• Create a list of capabilities.
• For each capability, include the importance (e.g., generally applicable across or only within certain projects/disciplines) and corresponding features needed to support the capability in a shared observation model.

Background

The need for a shared data model for environmental observations is motivated by the capabilities that such a model would support, and these capabilities will determine the degree of detail and formalization required in the model. Many different uses of a shared observational data model can be envisioned. The goal of this session is to identify needed capabilities, resolve the importance and potential for these capabilities, and to clarify the modeling requirements needed to enable these capabilities. Examples of capabilities that an observational data model could facilitate include: improved discovery of relevant observations across highly heterogeneous collections of data; clarification of semantic relationships among observations that might have been taken under varying data collection protocols or as part of differing experimental regimes; and the integration of observational data to facilitate new scientific analyses.

Discussion topics

What capabilities can be envisioned for a shared observational data model? What model features are needed to support these capabilities?

Possibly issues:

• What generic capabilities are needed across domains and projects?
• What specialized capabilities are needed for specific domains and projects?
• What challenges exist for the different capabilities? (e.g., data intensive, compute intensive, metadata intensive, etc)
• Do existing solutions (implementations) exist either partially or fully? Are they open source, extensible?
• For your project, what are the driving applications with respect to observational data?
• How should capabilities of a shared data model be prioritized? Do some depend on others?

Session notes and results