Mike,
One example of a problem would be that we'll need a representation of
dates and times. SUMO has that already, and uses a composition of
functions. It's an efficient and elegant way to represent this sort of
information in logic, but very ugly if one tries to convert it to a frame
representation. This is similar to the issue discussed on the telecon last
week that Pat brought up about representing units. Because this
information is better represented in a different way in a frame system,
this leaves us with either not having such information in the frame
version, or of wasting effort to recreate the portion of the ontology in a
frame-friendly format. Both options have a significant cost, so that's why
it makes sense for us to move from frames to logic very soon, or now.
In detail, if you wanted to represent that an order occurs today, we
could say (01)
(equal
(BeginFn
(WhenFn TodaysOrder))
(DayFn 6 (MonthFn August (YearFn 2003)))) (02)
A typical frame representation might be (03)
(dateOfStart TodaysOrder "8/6/2003") (04)
A better frame representation would be (05)
(dateOfStart TodaysOrder TodaysDate)
(day TodaysDate 6)
(month TodaysDate 8)
(year TodaysDate 2003) (06)
Now, one might say that the frame representation looks simpler, and it
does. However, it's fairly specialized because it requires metric
dates. The logical representation allows us to state that an event occurs
on a certain day of the month without specifying a year or month. It
allows us to relate the start of an event to a time, or even another
event. That flexibility allows us to state some general axioms such as if
A occurs before B and B before C then A occurs before C.
One can't just put off considering the axioms until later, because they
depend on the basic terms and relations. If the terms and relations are
created in a limited frame language, it will be make the axioms very hard
to state in a general way later.
If we agree that axioms are important to have (at some point in the
development process), then its important right now to create a
representation that will accommodate them later. (07)
Adam (08)
At 11:01 PM 8/6/2003 -0400, MDaconta@xxxxxxx wrote:
>Hi Adam,
>
>After looking at the reasons for KIF (trinary predicates, negation, and
>lack of axioms),
>it seems to me that all those cases you mention refer to axioms or rules
>which was
>to be step 8 of our process. I still don't see why protege cannot be used to
>represent the majority of our classes and properties (our facts) and then
>we move to KIF to implement our axioms and rules per Step 8. Also, those
>cases
>you mention that affect facts (trinary relations and negation) seem to be
>either
>corner cases or of indeterminate importance to our modeling.
>
>I am not opposed to KIF but I am not yet convinced that starting with
>frames biases
>the process in the wrong direction.
>
>- Mike
>
>
>In a message dated 8/6/03 9:47:38 AM, adampease@xxxxxxxxxxxxx writes:
>> I've posted my summary of the reasons we had for our original consensus
>>to use first order logic for the ontolog effort. I've also included the
>>text below.
>
>
>
>
>
>
>-------------------------------
>Michael C. Daconta
>Chief Scientist, APG, McDonald Bradley, Inc.
>www.daconta.net (09)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (010)
|