W3C home > Mailing lists > Public > www-rdf-interest@w3.org > February 2002

Re: Provenance in RDF

From: Dave Reynolds <der@hplb.hpl.hp.com>
Date: Wed, 27 Feb 2002 17:50:48 +0000
Message-ID: <3C7D1C78.D521B680@hplb.hpl.hp.com>
To: Charles McCathieNevile <charles@w3.org>
CC: "RDF Interest (E-mail)" <www-rdf-interest@w3.org>
Charles,

Thanks for these pointers. I hadn't realized till I read these that EARL is
using reification.

The idiom you use is interesting. This seems to be of the form:
   fred asserts {subject hasProperty x}.

Such a fact base on its own doesn't actually contain the triple "subject
hasProperty x". For the EARL application I can see that is very useful. For us
we actually regard all these assertions as being valid and are the primary
things we query, the provenance information is secondary. So if we were to take
the reification approach we'd probably use something like:
   subject hasProperty x.
   {subject hasProperty x} hasProvenance [:assertedBy fred ....].

Whether to make the reified statement the subject or object of the provenance
statement is probably not significant - just a matter of modeling choice. It
seems slightly more intuitive to attach rich annotations to the provenance
statement using bNodes rather than embedding them within the reification but
either way would work fine.

Thanks for your help.
Dave

Charles McCathieNevile wrote:
> 
> The EARL project http://www.w3.org/2001/03/earl (and other stuff from
> http://www.w3.org/WAI/ER for humans to follow) - deals in provenance.
> 
> There are schemas and some discussions that provide further explanation of
> how and why we did what we do, and there is some ongoing discussion about
> whether we can simplify what we have so far.
> 
> EARL is an evaluation reporting language that was built in the context of
> accessibility evaluations. One of the use cases is that different people or
> tools may have different ideas of how accessible a particular thing is (even
> to the extent of disagreeing whether it meets a fine-grained requirement).
> Another is that a resource may change, so having a time for the provenance is
> important too.
> 
> there are now a couple of different tools working with EARL, and we hope to
> have more soon, so we might not have all the answers but we are facing the
> same questions. (There are others on this list who have been working on it
> and maybe able to expand on this...)
> 
> cheers
> 
> Charles
> 
> On Wed, 27 Feb 2002, Dave Reynolds wrote:
> 
>   We are working on a semantic web related application that needs some provenance
>   support. We have various routes for doing this but would be interested in
>   hearing of other's experiences. Are there any groups out there that have
>   developed applications supporting provenance within RDF that would be willing to
>   share their experiences on what worked well or badly?
Received on Wednesday, 27 February 2002 12:51:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:53 GMT