Date

Attendees

Notes

  1. "API idea" decompose the loader into components to facilitate component level use
  2. "Distribute with product"
    1. Must be the way newcomers can start
  3. "Java eventually."  
    1. We can prototype in Python
    2. Functionality
    3. Modularity
  4. Three primary domains
    1. Orgs
    2. People
    3. Works
  5. Secondary domains
    1. Journals
    2. Dates
  6. URI creation, URI discovery.  Key fields, round trips involving the URI.
    1. Generated from the key field (see OpenVIVO).  Natural key identifier.
  7. Requirement entities loaded by the data loader must:
    1. be "updated" by the data loader
    2. participate with other agents regarding the entity.  Subgraphs managed by various agents.
  8. Identity integrity problems across the agents?
    1. Can we determine that the attributes to be managed by one agent have been changed by other agent?
  9. Converting to the data loader?
  10. VIVO needs an attribute on the Authorship:
    1. Confirmed – mine
    2. Confirmed – not mine
  11. VIVO needs an interface the authorship confirmations
  12. Update. "Day 2"
    1. "brain surgery on the kb-2 graph" – is this as difficult as we each find it?
    2. Named graph – restricts the triples that need to be managed, provenance, can drop and reproduce, self-edit turned off for attributes managed by the loader
  13. How to do the work
    1. Requirements
    2. Parameters – what can be said about what we want done?
      1. Minimum case.  Here's a list of people.  Discover the works. Create data.  Put it in.  Replace whatever we had about those people.
      2. Learning.  If its not your work, put it in the list of "not mine" and don't repeat the mistake.  If it is your work, add to the list of confirmed works to improve the hueristics for finding works and insure the work will come back.
      3. Email the person whose data has changed.  Selectable option.
    3. Large scale design / APIs
    4. Work down toward modules

 

1 Comment

  1. Prerequisite for No 10 should be something like a nightly changes report, something that allows an editorial staff to look after changes, especially regarding to the quality of imported data.