Re: Issue 3 in void-impl: provenance information

3 views
Skip to first unread message

void...@googlecode.com

unread,
Feb 22, 2011, 7:14:54 AM2/22/11
to void-di...@googlegroups.com

Comment #15 on issue 3 by tope.omi...@googlemail.com: provenance information
http://code.google.com/p/void-impl/issues/detail?id=3

I have been working on voidp, a provenance extension to void. It's here:
http://www.enakting.org/provenance/voidp/20110210.html . It is designed to
be lightweight. It'll be good if these can be incorporated into void.

void...@googlecode.com

unread,
Oct 13, 2012, 10:42:30 AM10/13/12
to void-di...@googlegroups.com

Comment #16 on issue 3 by rich...@cyganiak.de: provenance information
http://code.google.com/p/void-impl/issues/detail?id=3

With W3C's PROV Ontology getting closer to Recommendation, it's probably
one of the first places we should look at. As far as I understand,
typically, PROV is not used directly but through specializations that are
then mapped to PROV. The specializations define domain-specific entities,
roles, activities, and so on.

The PROV Working Group has a Dublin Core to PROV mapping, so this might
already solve part of the problem.

And VoID could define further activities and roles specific to RDF
publishing, perhaps inspired by voidp, and map those to PROV.

Reply all
Reply to author
Forward
0 new messages