RE: Annotations and entailments

No answers in this email, I'm afraid, but a couple more interesting
questions.  Sorry.

> From: Jeremy Carroll [mailto:jjc@hpl.hp.com] 
> FileA:
> <owl:Class rdf:ID="example">
>    <rdfs:comment>An explanation</rdfs:comment>
> </owl:Class>
> 
> 
> FileB:
> <owl:Class rdf:ID="example">
>    <rdfs:comment>A different explanation</rdfs:comment>
> </owl:Class>
> 
> 
> Currently
> 
> FileA DL-entails FileB
> 
> and
> 
> FileB DL-entails FileA
> 
> While being IMO simply incorrect; this also breaks semantic 
> layering since in OWL Full (like in RDF) neither entailment holds.

Is it 'incorrect', though?  Seems to me this is the same problem as
whether the following two files are 'identical':

FileA:
int main () { return 0; /* Do nothing */ }

FileB:
int main () { return 0; /* Successful termination */ }

There are at least three views of this:

1)  equivalence of code produced by running these through a C compiler
(should be identical output);

2) equivalence of the files via 'diff' (they're different); and

3) the intention of the author (unknown; comments appear to imply that
one is considering the program's operation whereas the other is
considering the program's effect on its environment).

Which of these views is/are 'correct'?  Which are 'simply incorrect'?
Ideally, I'd like to be able to choose my own interpretation between
view 1 and view 2 depending on whether I was running a compiler or a
source code control system.  I could live with just view 1 or just view
2 provided it was made very clear to me what was intended; I would be
very unhappy having to live with view 3.  Seems to me that the RDF spec
is currently somewhere between view 2 and view 3.  I don't actually know
what that means to me as a writer of OWL or of RDF.

Here's another interesting one, by the way, more related to
rdfms-assertion; I'm not sure what to make of this.  Consider a variant
of the UMD DAML+OIL to OWL translator that takes DAML+OIL and/or KRSS as
input and that produces OWL output.  Consider further that it is
accessible via a HTTP GET and can translate DAML+OIL and KRSS that it
can retrieve by URL - so there is a unique URL for its OWL translation
of some non-OWL (and, indeed, non-RDF) input.  Consider further that
some public-spirited soul makes this available as a service on their Web
site.  Who is responsible for the 'social meaning' of the produced (and
effectively published) OWL?  The author of the original document,
despite the fact that they wrote in a formalism that didn't have this
burden?  The operator of the translation service, despite the fact that
they have no control over the data on which it is used?

Taking this and the rdfms-assertion issue, the whole thing is a pretty
serious can of worms.

		- Peter

Received on Sunday, 19 January 2003 09:09:27 UTC