DatasetDynamics/Meetings

From W3C Wiki
Jump to: navigation, search

This page is for collecting topics and notes for Dataset Dynamics meet-ups.

Kick-off Meeting at WWW2010

Proposed venue: LOD W3C Track at WWW2010

Time: 3:30pm

Organiser: Michael Hausenblas (remote only: mhausenblas on IRC, skype, Twitter) and Juegen Umbrich (local organisation)

IRC back-channel: #swig on Freenode (see also Semantic Web Interest Group IRC Scratchpad)

Participants:

Topics

  • Change detection in LOD datasets (algorithms, methods, etc.)
  • Lessons learned (e.g., RPI's work on data.gov)
  • Discovery (dady, etc.)
  • Notification mechanisms (Atom, etc.)
  • Delta representation (GUO, Talis Changeset, etc.)
  • Application areas (example Use Cases)
  • Possible standardisation activities (W3C XG?)
  • Dataset Dynamics Testbed (Benchmark datasets)

Notes

Use cases

(ordered by complexity)

  • Smart caching
  • link maintenance
  • dataset synchronization

Granularity

  • resource level
  • triple/quad level
  • datasets

Communication protocols

  • Atom
  • OAI-PMH

Communicating Diffs?

  • not all data providers may support that
  • Talis changeset vocabulary for expressing low-level changes

What has to be communicated?

  • notification of change
  • diffs (deltas), enough information to mirror the change
  • WHY something has changed, change metadata, changes as first-class objects with URIs

Architctures

  • scalability is THE issue
  • Publish/subscribe
  • batch changes
  • combinations

Describing dataset dynamics

  • datasources should dscribe their own dynamics
  • Change frequency
  • pointers to feeds, etc.
  • extend VoiD for that