Re: [TF:DbE] The easiest keys there are

I Am Not A Logician...

Imagine that we had keys that are strings consisting of 3 numerals (0..9). 
Let's say we also have keys that are strings consisting of 2 hex numerals 
(0..9,a..f), and lastly that all instances of a single class have exactly one 
numeral key and one hex key (independently assigned).

The first key forces there to be a maximum of 1000 instances of this class. 
The seccond key forces there to be a maximum of 256 instances of this class. 
This is a direct result of the keys being mapped injectively on to the 
instances.

Now, given 200 asserted instances, some of which have asserted values for the 
numerical key, some for the hex key and some for both, what should we do? 
Well, with the strong semantics, we should go about trying to infer which 
individuals may be the same based upon their key values *while symultaneously 
infering the value of the non-asserted keys*. So, the reasoner now has to try 
a possible world matching each and every way of handing out the 1000 * 256 
combinations of keys to the asserted individuals and potentially also to 
individuals that are inferred to exist due to existential and cardinality 
restrictions. It is only able to make conclusions about the consistency of 
the ontology and then from this, the most specific class(es) of each instance 
and values of the key properties by finding the axioms that are consistend 
with every one of these expansions. This search is painful. It gets worse as 
the key space expands, as long as it remains bounded.

Is that about right Bijan?

Matthew

On Wednesday 03 October 2007, ewallace@cme.nist.gov wrote:
> Back to the using keys with OWL discussion:
>
> Bijan,
>
> I infer from what you wrote and a quick look at the
> poster that you have concluded that keys as integrity
> constraints with OWL are problematic.  Can you elaborate
> on this in a fashion that a non-logician might understand?
>
> -Evan
>
> Evan K. Wallace
> Manufacturing Systems Integration Division
> NIST

Received on Wednesday, 3 October 2007 14:28:53 UTC