Re: new names for OWL lite/fast/large

>    OWL professional or OWL pro

> and we can remain silent about the target profession for the
> last.



More seriously, I think I am with Chris in prefering
  "OWL" := "large OWL"

I think that we should define conformance, and that for reasoners possible
conformance is quite complicated. (As already noted in this thread).

So I can imagine a product described as:

- an *OWL Reasoner*
   supporting *OWL DL*
   with *classes as instances*
   and *distinguishing typed values*

or somesuch, where the starred terms are defined in our specs; and in which
the starred terms relate to test cases that must pass.

We will need to be careful about the complete/incomplete thing.
We clearly need to permit incomplete reasoners at some level.
A complete fast OWL reasoner, would be incomplete if regarded as a large OWL
reasoner. Marketing people will be against a conformance level with the word
"incomplete" in, so we perhaps need to identify what people need to do to
use the term "complete ?? OWL ?? reasoner". I don't think anyone is
expecting a complete large OWL reasoner, and so we should perhaps drop the
adjective, except for reasoners which don't meet some target - these could
then (perjoratively be) "incomplete OWL reasoners". We could set a low bar
for OWL Lite reasoners, again with no adjectives except for a perjorative
'incomplete' for something that really can't do much at all.

With the fast OWL level, it makes sense (at least to me) to define what it
means to be "complete" meaning complete with respect to the syntactic
restrictions implied in the formal semantics.
On such a model we could either have a perjorative "incomplete" level
defined lower than a technically incomplete level - or insist on a
technically correct use of incomplete.


Received on Monday, 21 October 2002 03:44:44 UTC