IW Meeting 2012-02-02

From Inference Web

Jump to: navigation, search

Contents

Meeting Information

Agenda

Discussion

prov-wg

  • Entities are _gone_, we're back to Things.

prov-wg is splitting "scruffy" modeling from "proper provenance" modeling.

  • Result is a Radically Reduced Model (TM) with an extra Proper Provenance (ProP) model.

The day ended a bit sour when we talked about prov-o. It's stagnant, nobody wanted to say anything, so Tim aired the laundry

  • the "catch up" process is unrealistic,
  • RDF as a second class citizen in a W3C wg is backwards.
  • we don't have quantitative measures to know how complete prov-o is w.r.t DM,
  • we don't get feedback from the broader wg (apparently b/c we didn't ask).

way forward:

  • make prov-o sync with latest DM WD (today's) (ONLY the ontology b/c wordsmithing will slow us down)
  • prov-o team notifies group and will get "coverage" feedback.
  • subsequent changes are considered w.r.t how they affect DM _and_ PROV-O
  • Paul will act as intermediary between dm and prov-o.

Paulo

held a provenance planning meeting (10 people) VERY diverse set of disciplines.

csv2rdf4lod

(Paulo brought it up, I swear)

showing a csv2rdf4lod provenance in Probe-It. TODO: tim to point the provenance to Hugo, Nick, Paulo.

the provenance of the process that csv2rdf4lod fulfills is ill-defined because it is not constrained to particular uses.

  • we have no upstream users or control of their processes.
  • we don't have downstream users (though, _this_ is possible).

problem that csv2rdf4lod's provenance addresses:

random sketchy student at RPI is providing you EPA data that is not in a form that the EPA provided - how can you trust him?

Tim asked Paulo to help him write the csv2rdf4lod-automation provenance paper.

Facts about IW Meeting 2012-02-02RDF feed
Date26 February 2012  +
Personal tools
Navigation