I attach a reply to Chris Mungall's from our man in OWL 2. (01)
--
Edward J. Barkmeyer
Manufacturing Systems Integration Division
National Institute of Standards and Technology
100 Bureau Drive M/S 8263 email: edbark@xxxxxxxx
Gaithersburg, MD 20899-8263 tel: +1-301-975-3528 (02)
--- Begin Message ---
Indeed, Chris Mungall is slightly confused, but he certainly can be
forgiven for it.
Unfortunately, his subtle misunderstanding is exactly about the aspects
of the
Semantic Web languages that are key to this discussion. (01)
> I'm not sure Ian is 100% wrong. Or if he is wrong, then he is at least
> in good company. Where OWL is concerned, it's common to treat RDF as a
> syntax - perhaps just because it simplifies things tremendously. See
> for example:
>
> "OWL 2 Web Ontology Language: Mapping to RDF Graphs"
> http://www.w3.org/TR/2008/WD-owl2-mapping-to-rdf-20081202/
> "This document defines a mapping of OWL 2 ontology into the RDF
> **syntax**, and vice versa. " (02)
Yes. Use of the word "syntax" in the last sentence in the abstract for
this document
is unfortunate. Of course, the DL'ers who authored the document didn't mean
concrete syntax, they meant a more abstract use of the term. But Chris
ignores
the part of the title that says "to RDF Graphs" and either didn't read or
didn't understand the remainder of the document. The mapping is to the
abstract syntax of RDF (RDF Graphs). With this, OWL ontologies and
facts can be stored in an RDF triple store or exchanged with any of its
concrete syntaxes (such as rdf/xml, Turtle, or corresponding n-triples). (03)
OWL DL and RDF semantics do differ, but Web Ont stood on its head
to make the RDF(S) interpretation of a set of RDF graphs describing OWL
DL 1 consistent with the DL interpretation. "Syntactic" restrictions
describing which RDF graphs could be interpreted as OWL DL almost
made this work. (04)
OWL 2 adds new features to the language that are harder to represent as
RDF graphs and still maintain a consistent interpretation. The mapping uses
the dreaded reification vocabulary to handle this. While the RDF language
includes this vocabulary, no semantics are assigned to it. Thus there
are additional
things in OWL 2 that aren't understood by RDF tools by the RDF semantics
alone,
but at least they are not misunderstood to mean something altogether
different.
>
> You'll have to forgive us mere mortals for getting confused here; it
> goes on to state "The syntax for triples used in this document is the
> one used in the RDF Semantics [RDF Semantics]" - which links to your
> document, which uses N-triples as the syntax. So is RDF a syntax? Why
> doesn't the OWL document say "defines a mapping of OWL 2 ontology into
> N-triples syntax"?
Because they didn't define a mapping to N-triples syntax. They simply used
N-triples syntax to describe the forms of RDF graphs being mapped.
>
> There is a separate document for mapping to RDF semantics:
> http://www.w3.org/TR/2008/WD-owl2-rdf-based-semantics-20081202/
> "This document provides the RDF-compatible model-theoretic semantics
> for OWL 2, called "OWL 2 Full"
>
The OWL 2 Full semantics is an RDF compatible semantics for OWL 2. It
is not a mapping to RDF. An RDF(S) interpretation of the graphs for OWL
only gives a partial understanding of its meaning. OWL Full semantics
extends the RDF(S) semantics to define the interpretation that OWL Full
tools should apply to OWL ontologies and data.
> But this can be ignored if one simply wants to use RDF/XML as a
> transport layer for OWL (which I'm guessing will become less common
> now there is a sensible XML syntax for OWL)
>
By "transport layer" I guess he means exchange syntax, and by "OWL"
I guess he means OWL DL. Actually, if one is only interested in OWL DL
or one of its profiles, one can safely ignore the OWL Full semantics. In
fact, unless one is implementing a reasoner, there should be little need to
look at any OWL semantics document.
> Now obviously the story is a little more complicated here, and it
> seems that the division between syntax and semantics is not entirely
> agreed upon. But there is nothing to stop Ian defining a similar
> mapping for his language, performing tricks such as making rdf:type
> triples disappear into the ether. I guess you could say that the OWL
> folks are "using RDF syntax (particularly rdf:type) in some other non-
> RDF way" - if so, is that bad? (05)
RDF folk think that this is bad, and the confusion created is certainly not
good. Not sure what allusions are intended by the mention of rdf:type,
having missed that part of the thread I guess. However, the requirement
for certain type triples is one of the "syntactic restrictions" on graphs
representing OWL DL. (06)
>
> Personally I don't really see the point, but I just thought I'd point
> out that Ian isn't in the minority here. (07)
The point would be to make use of RDF tools for parsing, storing,
querying, etc. This is particularly helpful if one wants to link the
content in one's language to other RDF based content. RDF has some
nice properties for distributing content, filtering, merging and
re-analyzing resulting graphs. To do this though, one would have to
understand enough about the RDF model so as to not be surprised by
the results. It sounds as though this fellow Ian may not understand this
model well enough (although there really isn't enough information in
this email to say). As I have said before, a meta-model inferred from
the structure of rdf/xml will surely mismatch the behavior of RDF
tools. Let's hope he is, at least, *not* making that mistake.
>
> Cheers
> Chris
-Evan (08)
--- End Message ---
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (01)
|