IW Meeting 2011-02-15

From Inference Web

Jump to: navigation, search

Note: IW Meeting 2011-02-15 was rescheduled to IW Meeting 2011-02-16.


Meeting info


  • Tim
  • Jitin
  • Nick


  • Jitin review his crawler work
  • Tim's plunk diagram
  • Nick update on "actions" for ProbeIt
  • Tim FOO ontology, Mapping encoding


Jitin's crawler work

Jitin's Objective (when he did this): how to query the PML collected by Trust Lab. includes SPARQL ontology


accessing files from different projects via public interface.

2 parts:

1) wget script that grabs ALL files starting at roots. needed to grab ALL b/c file extensions were not consistent.

  • did NOT retrieve by file extension, did simple exclusion (pdf, xls).
  • why does it download the thing you said not to? b/c it needs to look into it for links out to other files on the server.
  • shell script has lists of roots and invokes wget.
  • DONE: Jitin to email shell script.
  • TODO: Tim check out, incorporate.
  • NOTE: files on server that ARE rdf without the normal extensions.

2) groovy script to look into file, figure out if it is RDF. If it is, load into local TDB directory.

handful of services:

Tim's version


  • encoding where PML files within PML (RDF).
  • Characterize existing "where is the PML"
  • Offering them to the public via a version control repository.
  • Demonstrating how to act upon the descriptions of where the PML docs are.

Jitin: what about just following the links?

  • Start dereferencing the PML instances from the seed of what you aggregated.
  • ProbeIt does the crawling.
  • The wget would do the instance-following within the scope of the server.
  • would miss inter-server linkages.
  • TODO: what is the inter-connectivity of PML instances across domain names?
  • e.g. iw.cs.utep.edu and trust.utep.edu inference-web.org/registery

Jitin: second pass on document seed to resovle the instances for linked data. Nick: now crawling some of the instances (e.g. owl vocab)


wants to show action as well as (use?) incubator group.

construct, {ignored, recognized, required} THEN action tool does. THEN what does the action do for the user THEN use case (can come from this direction, too).



Personal tools