W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > March 2013

Re: owl:sameAs - Harmful to provenance?

From: Jim McCusker <mccusj@rpi.edu>
Date: Thu, 28 Mar 2013 00:08:09 -0400
Message-ID: <CAAtgn=RAC62icWML5qWCcQPQ2iX+bhV6jtmMw=x2BuYzVntHLg@mail.gmail.com>
To: Pat Hayes <phayes@ihmc.us>
Cc: Rafael Richards <rafaelrichards@jhu.edu>, Oliver Ruebenacker <curoli@gmail.com>, David Booth <david@dbooth.org>, "<public-semweb-lifesci@w3.org>" <public-semweb-lifesci@w3.org>
On Wed, Mar 27, 2013 at 10:56 PM, Pat Hayes <phayes@ihmc.us> wrote:

> Well, I agree that is a problem, but don't draw the conclusion that there
> is something wrong with sameAs, just because people keep using it wrong. I
> guess your analogy to GOTO put my back up a little.

The analogy to GOTO was pretty apt, actually. There's nothing wrong,
inherently, with GOTO. Both are necessary to implement higher-level
functions (inverse functional, max cardinality, etc.), both are/were
misused in ways that tended to introduce "bugs" (incorrect inferences), and
both should be very carefully considered before they are used directly. We
selected that title to provoke discussion (it was an abstract submission),
which it successfully did.

> Well, I'm not (yet) sure what "entity provenance" means, but the main
> point is that if you have some information about A, and A owl:sameAs B,
> then that information is also about B (because A and B are the same thing.)
> I was assuming that provenance was not about the object, but if it is, then
> sameAs should apply to it in the same way as it does to any other assertion.

This is true, unless you need to do something that needs to distinguish
things (HeLa cell sample 12345 was used in a gene expression experiment)
and then not distinguish them (TERT is upregulated in HeLa). Entity
provenance is the history of an entity, and follows all of the same
derivation rules and events that data provenance might, but it's about the
entity itself. One modeling challenge is expressing changes of state in
entities, such as freezing a tissue sample, without changing it's essential
nature (it's still the same sample, just different things are true about it
now). prov:specializationOf and prov:alternateOf allow for granular
expression of these sorts of things without getting mixed up with sameAs or
requiring that you stop denoting the thing and instead denote some role of
the thing.

> But OK, I will have to go read your stuff before commenting further, and
> that might take a few days.

My IPAW 2010 paper [1] motivates entity provenance best, even if the
application is outdated.

[1] McCusker, J. and McGuinness, D.L. 2010. Provenance of High Throughput
Biomedical Experiments. In Proceedings of 3rd International Provenance and
Annotation Workshop (June 15-16 2010)

Jim McCusker
Programmer Analyst
Krauthammer Lab, Pathology Informatics
Yale School of Medicine
james.mccusker@yale.edu | (203) 785-4436

PhD Student
Tetherless World Constellation
Rensselaer Polytechnic Institute
Received on Thursday, 28 March 2013 04:08:53 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:53:02 UTC