On Sonntag, 8. Juni 2008, Jeff Thompson wrote:
> At the beginning of SMW development, the idea was that, to make everyone
> happy, anyone could create any property that they want and give it whatever
> meaning they want in their wiki.  That's great for making everyone feel
> included (even the wiki users who just want to jump in without thinking
> about what they are doing).
>
> But as soon as SMW has to do inference, even the basic inference for
> transitive and inverse properties, then this requires not only a
> syntax, but also a semantics.  Some people don't like (or don't understand)
> the semantics developed for OWL, but at least it is well-defined.
>
> Does SMW use the semantics defined for RDF/OWL by default?  If a wiki
> user wants a different semantics (for what defines equivalent topics, etc.)
> does SMW support a language for them to define it (pretty difficult)?
>
> You see where I'm going with this question.  Doesn't SMW need to get
> past the "anything goes" and officially commit to some well-developed
> semantics such as OWL?

I see I need to clarify some things here: SMW certainly *is* committed to the 
OWL DL semantics. As you may know, I am also member of the current OWL 
working group at W3C, and SMW certainly will pick up some results from these 
new developments around OWL.

Now you refer to the option to "create any property" that someone wants. This 
is no conflict with OWL: SMW properties are just Object- or 
DatatypeProperties (depending on SMW datatype), and these can of course be 
created by users (just like in OWL). But there is no inference involved 
there. Of course, we cannot prevent someone making a property 
called "inverse" or a category called "TransitiveProperty". But this then is 
just a confusing naming that reminds humans of semantic (OWL) features 
without having any formal meaning for SMW.

The reasoning constructs currently defined for SMW are very limited: basically 
there are only class and property hierarchies (with the special case of class 
and property equivalence), and some equality statements. The implementation 
of those corresponds to the official OWL semantics. Everything else is just 
assertional knowledge (data, "Abox axioms"). You can also see the official 
OWL semantics in SMW by looking at the OWL/RDF export. In principle, OWL 
queries against that export should result in the same results as if executed 
inside the wiki. By the way: SMW #ask queries can be described by suitable 
OWL DL class expressions, so SMW querying also has a clear mapping to OWL. As 
I indicated before, we also are going to further extend the reasoning 
features provided by SMW, and these extensions will also relate to the 
expressive means available in OWL (2) DL.

I know that we have discussed here various ways of implementing inverse 
properties, but each of those appears to be compatible with OWL DL. 
Unfortunately, inverses in SMW must be confined to "named individuals" if SMW 
should ever include the power to assert existential knowledge -- otherwise 
complexity would blow up to ExpTime[1], which is certainly not feasible in a 
large wiki. In the OWL literature, such restricted axioms have been studied 
under the label "DL-safe rules", and in our case we can imagine them as 
a "macro" that makes SMW add annotations to complete the user input. Was your 
concern related to the use of the name "inverse" for something that is not 
the OWL inverse?

So at which place do you think SMW is not officially committed to OWL 
semantics? We usually avoid bothering users with the details of OWL DL, in 
particular since only a few OWL constructs currently occur in SMW. Should 
there be more documentation on the relationship to OWL DL?

-- Markus

[1] This says: the proper reasoning in principle is very very hard to 
implement, and only by algorithms that use a rapidly growing number of 
computing resources.

>
> - Jeff
>
> Markus Krötzsch wrote:
> > Yes, the only possible *semantically distinct* inverse is P (which is
> > always equal to R in this example). But some wiki user can still make two
> > distinct property pages for P and R, and SMW then has to figure out that
> > these are equivalent *for the current information on inverses*. If the
> > inverse-assertions in the wiki change, then statements for P and R may
> > again become different from each other, so SMW would also have to keep
> > them separate in some sense.
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://sourceforge.net/services/buy/index.php
> _______________________________________________
> Semediawiki-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/semediawiki-devel



-- 
Markus Krötzsch
Semantic MediaWiki    http://semantic-mediawiki.org
http://korrekt.org    [EMAIL PROTECTED]

Attachment: signature.asc
Description: This is a digitally signed message part.

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Semediawiki-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/semediawiki-devel

Reply via email to