We are trying to tie together all of the artifacts related to the provenance and context content around data produced by a project or mission. For our model to work, we need to recommend identifiers for all those artifacts. Following the principles of linked data, we need to map all identifiers into URIs to we can describe relationships between them and assert facts about them.
- Artifact identifier characteristics:
- Unique - within our model, each artifact can still have other identifiers
- Must be able to map into a URI for linked data
- What does that mean?
- Field Experiment
- See Data Identifiers
- Must always have a unique identifier
- UUID for easily minting unique identifiers
- EOS 'LocalGranuleID' (filename)
- "provenance equivalence identifier"
- DOIs resolving to a canonical web 'landing page' for the dataset as a whole
- "Unique set of specific granules"
- "Equivalent dataset"
- Distinguish versions of these
- Includes 'calibration software' and 'product software'
- Algorithm -- The 'general' algorithm
- Algorithm software source code (DAP?)
- Algorithm executable package
- Other non-algorithm software?
- A bunch of the PCCS artifacts are mission level 'documents'
- Scientific literature related to the mission/project/dataset/etc.
- Can of worms... Back to Bruce's nomenclature exercise..
- Need both for 'granule' level and 'collection' level
- Role(s) in the project