- From: Timothy Lebo <lebot@rpi.edu>
- Date: Sun, 15 Apr 2012 15:03:21 -0400
- To: Jun Zhao <jun.zhao@zoo.ox.ac.uk>
- Cc: public-prov-wg@w3.org
On Apr 12, 2012, at 12:51 PM, Jun Zhao wrote: > Curl, > >> >> The "same thing" isn't always well-defined. Is the wood the "same >> thing" as the chair? >> >> In my example, file x and y have a significant number of AV pairs in >> common. Is that enough that for some definition of "same thing", they >> are indeed the same thing? >> >> >> I like these definitions for their simplicity and the almost >> mechanical way you can determine them. They just depend on the >> definition of "same thing". If we leave that open to the provenance >> describer to determine, this could accomodate whatever use cases were >> needed. > > I totally agree that we should really let the provenance providers to decide by themselves when they want to determine two things are same thing. > > I have thought this is the beauty of our model. And whatever definitions we have in the end, it should still permit providers to do so, and not confuse them or restrict them. It's the mechanism that we are interested in providing, not a define distinction for a very philosophical topic:) +1 specOf/altOf is the bridge that lets one add order to a combination of assertions that came from a "scurfy" (not distinguishing differences) asserter and a "proper" (distinguishing differences) asserter about common _t_hings. (is there a way to make that even LESS capitalized :-) -Tim > > -- Jun >> >> In my example, I could make that call and assert that since they were >> produced by an equivalent application of a reproducible process, the >> files should be equivalent and therefore specializationsOf some ideal >> thing and therefore x is alternateOf y. >> >> Curt >> > > >
Received on Sunday, 15 April 2012 19:03:51 UTC