Dial In Details
Date: Friday August 28, 2pm EDT (-4 UTC)
- U.S.A/Canada toll free: 866-740-1260, participant code: 2257295
- International toll free: http://www.readytalk.com/intl
- Use the above link and input 2257295 and the country you are calling from to get your country's toll-free dial-in number
- Once on the call, enter participant code 2257295
Meeting Goals
To review and refine (if necessary) use cases
To assess whether or not we have enough use cases
Determine next steps
Attendees
- David Wilcox
- Stefano Cossu
- Tim DiLauro
- Elliot Metsger
- Daniel Davis
- Aaron Birkland
- Ruth Duerr
- Kevin S. Clarke
- Andrew Woods
Agenda
Discuss Use Cases
Are they adequately described?
Do we have all the use cases we need or want? Are there more?
Process for reviewing and accepting use cases, scope, and requirements
Do we vote on them?
Prioritize them in some way?
- Next Steps
- Any Other Business
Related Resources
Design Page (with use cases outline)
Previous meeting agenda, including minutes
Minutes
- Put agenda on Fedora Community (seems OK for everyone)
- Working down the list of use cases
- Stefano - Content Model
- Content modeling
- Have a model resource
- Use content model (with parameters) to formulate the form of the data returned
- Use to make/handle derivatives, metadata extraction, feature extraction
- Not hard code content model but use the model to guide processes
- Manage types and control relationships and other controls (cardinality, unique IDs etc.)
- Manage inheritance of modeled features
- Andrew, draw out common patterns. Where is the starting point?
- Andrew, implies complex step, multiple components ingest, Sparql, converters, validation
- Aaron, described possible common elements
- Stefano, model concentrates on the wiring dispatching to the right tool
- Aaron, we need a clear scenario that illustrates an example how things should work
- Stefano, started an implementation but the scope was too big
- Aaron, an example may use a set of object paths
- Stefano - Access Controls
- Would cover more components such as the triple-store with a comprehensive access controls
- Ruth - Federated Search + more
- Researchers want to be able to search over many repositories
- OpenSearch informs this
- Do layers of search
- Classified searches for the kind of data (discipline)
- Elliot - more than Federated Search (what is amenable to a kind of search)
- Elliot - How should we proceed
- Aaron - Extract common patterns
- Stephano - Use a new section to describe common patterns
- Andrew - Hoping that we would have more understanding of use cases for discussion and identifying patterns
- Elliot - Can we try to get the Wiki page for patterns in one week so we can discuss in two weeks (Wiki deadline - Friday Sep 4)
- Elliot - How can we communicate especially questions? Use comments on page?
- Tim - Wiki page better because it too easy to lose track on email
- Elliot - IRC or Slack or both, what channel is good for folks?
1 Comment
Stefano Cossu
This is the project that I mentioned during the meeting: https://github.com/aic-collections/aicdams-sspad
It is a service provider that is meant to achieve what I have described in my use cases: i.e. an abstraction layer for index search, CRUD, authN/Z, content modeling, validation and pre- and post-processing of resources.
See specifically:
Controller assigning actions for a content model: https://github.com/aic-collections/aicdams-sspad/blob/test/sspad/controllers/asset_ctrl.py
Model defining services and their configurations for each action: https://github.com/aic-collections/aicdams-sspad/blob/test/sspad/models/static_image.py
Abstraction layer for Fedora repo: https://github.com/aic-collections/aicdams-sspad/blob/test/sspad/connectors/lake_connector.py
Abstraction layer for triplestore index: https://github.com/aic-collections/aicdams-sspad/blob/test/sspad/connectors/tstore_connector.py
The only feature that is not contemplated here is the configuration- and wiring-based content modeling (models are hard-coded Python classes which define methods and services for each content model).
As I mentioned, the project got too big to be managed by our small team and our tight timeline; but I think it might be useful for understanding my end goals.