External review

Hello,

I pass on some comments from an external reviewer with interest in provenance in clinical research and workflow settings, Vasa Curcin of Imperial College London. The versions he reviewed are, I believe, those available at the end of May.

"I have finally gone through the PROV documents, and I quite like the revisions/derivations and the addition of plans... However, I am slightly confused by the introduction of alternate entities and specializations. In a way, this is adding an inheritance structure to provenance entities, but in doing so, does it not go against the grain of the core idea - namely that a record is kept of actual chronology of events that occurred?

The example you give in the PROV Primer is very indicative of the issue, whereby a blogger refers to an article, and the website where the article is located later redirects the link to another version of the article. This action, in a way, changes the actual record of what happened, since it is not known anymore that the blog post had previously referred to the other version of the article. While one could argue that this ensures that the provenance link points to usable information, any audit of the process would be invalid, since the original information was lost. Imagine that the blog post was criticizing the article - once the provenance link has changed, this becomes incorrect. While I agree that capturing the notion that several entities are identical in some way is neccessary, maybe inheritance is not the best way to go about it."
Thanks,
Simon

Dr Simon Miles
Senior Lecturer, Department of Informatics
Kings College London, WC2R 2LS, UK
+44 (0)20 7848 1166

Requirements for Provenance on the Web:
http://eprints.dcs.kcl.ac.uk/1399/