Details
-
Architectural Decision
-
Resolution: Done
-
None
-
None
Description
Why now? We will likely start generating some initial data products at AA0.5, and will need to start tracking test data even before that.
Qualities: Modifiability, maintainability, (performance?), usability
As we are (ever so slowly) moving towards having common solutions for data management, one problem that is going to come up is how we keep multiple kinds of data without eventually losing the context that enables us to work with it. Therefore we should likely try to come up at least with at least some initial standards to indicate for instance:
- What kind of data it is (include a way to indicate schema, physical data model...?)
- When and where the data was recorded
- Who created the data (including what software [versions] were used)
- For what purpose it was recorded
- Any other relevant context data
This sort of information should likely be added to telescope configuration strings in a way that different sub-systems (but chiefly SDP) can use this to tag generate data.
