20 July 2017 Meeting
1. Orient attendees
2. Discuss the difference between the PID kernel information of a data object and its metadata.
There is tendency to view PID kernel information as a store for metadata. This could be because people view the PID system (or DOI system) as a sustainable place to store metadata; more sustainable than a metadata server for instance.
Discussion illuminated several useful points of view for thinking about content of kernel information:
a. "PID centric view", see https://www.rd-alliance.org/group/data-fabric-ig/wiki/df-configuration-p... - it's the last figure on that page
b. Motivating use case of versioning and traceability
c. Kernel information as it would augment the DOI system. DOI system does not support kernel information at present. It maintains metadata in a separate server accessible through RESTful and OAI PMH interfaces. Kernel information could be viewed as cached information from DOI metadata that gets stored at the local handle server. In this view, the kernel information is selected based on the need for fast access to certain information.
2. Discuss open questions from P9: https://www.rd-alliance.org/group/pid-kernel-information-wg/wiki/plenary.... Didn't get to this so will discuss in next meeting.
3. Open issues. None. Next meeting 3 Aug at same time 09:00 EDT / 13:00 UTC