Search Mailing List Archives
[protege-owl] Query: Subclasses or Sibling
tredmond at stanford.edu
Tue Apr 3 10:23:06 PDT 2007
On Apr 3, 2007, at 9:03 AM, Thomas Russ wrote:
> On Apr 3, 2007, at 2:59 AM, william fitzgerald wrote:
>> Method 2:
>> I think dealing with individuals is more a rules engine thing than
>> a DL
>> reasoner area.
> That depends on the particular DL reasoner.
> Other DL languages and reasoners have had mechanisms that are quite
> capable of reasoning using combinations of closed world reasoning,
> limited closed world reasoning, and even cardinality restrictions
> asserted directly about instances.
This thread has been going on for some time so I may be missing some
context. But it seems to me that you are making a link between
reasoning about individuals and closed world reasoning. I think that
these are two entirely different issues.
Ian Horrocks, Ulrike Sattler and many others have put a tremendous
amount of work into finding a reasonable computable description
logic. There results include algorithms for deciding questions about
individuals as well as classes. Pellet and Fact use the algorithms
that they have found so these algorithms are able to do "full OWL
inference" for individuals. (I am always tempted to use the word
"complete" here but I don't know if this is one of its technical
Now when you start talking about closed world reasoning, you are
talking about a different logic than OWL DL (and even than OWL
1.1). There are some suggestions flying around about how OWL can be
extended to handle the closed world. Two that I know of are
autoepistemic logics and default logics. There are also
implementations for these. I believe pellet has an implementation of
an autoepistemic logic and there was a paper recently about an
implementation for a default logic.
More information about the protege-owl