[OWL] [1.1] Second version of OWL 1.1 feature list
Ulrike.Sattler at manchester.ac.uk
Thu Dec 8 17:50:03 EST 2005
On 8 Dec 2005, at 14:32, Bijan Parsia wrote:
> On Dec 8, 2005, at 2:17 PM, Uli Sattler wrote:
>> a) admissible and
>> b) for which we can find a "conforming" syntax
>> Point (a) is clearly specified, and it ensures that reasoning
>> remains decidable and that it can always be done "in the same
>> way", regardless of the specific dataype in question: simply (?)
>> plug in a datatype reasoner for your datatype into the OWL-
>> reasoner, and you have an OWL-reasoner that can handle your datatype.
> We should separate these two points.
sure - this is why I made them ;)
> As long as the reasoning can be done in this oracular manner, why
> not give the users the rope to hang themselves? (I would suggest
> rather fine grained checking would be required, but even dangerous
> features could be used safely and detectible so.)
I am not sure I know what you mean...do you mean that we should allow
all kinds of datatypes, as long as they are admissible/oraclable?
This would be great -- but how can we "fix the meaning" of a
datatype? I.e., can I "export"/'describe" my datatype? or do I simply
ship a datatype-reasoner with an ontology using a custom-built datatype?
> OWL mailing list
> OWL at lists.mindswap.org
More information about the OWL