been watching this thread on standard ontologies for a bit now…I even
dared stick my head in the >lion’s
mouth a couple of times by posting.
realise the thread is supposed to be about using ontologies in standards.
Although that doesn’t >preclude
everyone doing their own thing (it certainly hasn’t before in the world
of standards), it would >perhaps
make sense if these standards *could*
all descend from a common foundation. By >foundation, I mean what Rich was calling
an Ultra-High-Level Ontology – i.e. ontic categories and >relationships.
anyone looked into the possibility of a foundation that covers the whole gamut
? First order, >higher
order, intensional, extensional (my sense), etc. ? RDF/RDFS gives us classes,
type-instance, and >sub-super.
It just doesn’t have individuals (my sense, ”something with
spatio-temporal extent”). RDFS >covers
higher order and first order by dint of not restricting how RDF:type is used.
All it lacks is a >grounding
in physical extent, which is understandable as it was designed to refer to
stuff in cyberspace >rather
than stuff I can kick.
One reason I liked your fundamental
ontology is that it is relatively compact, which I think is excellent for a
UHLO FO. Although the “whole gamut” is needed ultimately, I
think the UHLO level should be slim and only model what people can agree
on. It appears that John Sowa, Pat Hayes and you still have differences
with the concepts in your model, and Ed Barkmeyer wants axioms in addition to
concepts before he is willing to call it an “ontology”. Personally,
I think axioms would be better added in ontologies below the FO seed ontology,
since I believe defining the axioms will be a difficult process to reach
I think we need two
tracks. The first is what we can all (most?) agree on, perhaps your FO –
that is what I’ve called a UHLO for lack of a better name. The “whole
gamut” can descend from whatever the UHLO is in gobs and bursts as
specific needs are announced and met. The pieces of the “whole
gamut” might be developed by adding concepts which (probably) fewer will
agree on, but which there are groups of agreement due to similar mission
But like you, I think a
very slim highest level ontology would add some momentum to the work of this
list by letting us have at least a few concepts in agreement among us. If
you want to modify your UHLO to meet the slimmest of widely agreed concepts,
which process could make it fit the need very well. Additions should be
those which can be debated and ultimately agreed to by various groups of
The result MIGHT
initialize John Sowa’s lattice of theories like a coat rack for later
ontologies. That could make his idea of a registry of ontologies come to
life. Progress made after the UHLO could extend the lattice in areas
needed to continue development. The infinite lattice might be further
away, but at least we could point to progress in agreeing on and
realizing some common structures.
I have no dog in this
fight. Any ontology might be the UHLO – I just want SOMETHING we
can agree on because it is small enough to be agreeable to (nearly) all of
Rich AT EnglishLogicKernel DOT com
difficulty, I suspect is in the intensional vs extensional bit. In
BORO/ISO15926/IDEAS, we would >have
one class with many names (e.g. “Equiangular Triangle”,
“Equilateral Triangle”), whereas the >intensional folks would have two classes,
perhaps with an equivalence relationship between them. >Has anyone analysed the nature of these
two approaches and come up with a common foundation for >both, or are they so fundamentally different
it can’t be done ? Surely some logician or philosopher >must have cracked this ?
resources, at one time, had class_by_intension and class_by_extension in them,
not sure if >there
was anything to join them up though. I think David Leal did the work on this.
>T: +44 207 193 4605
>M: +44 7768 892362
Futures Limited is a company registered in England
with company number 05248454
Company Address: 1 Nelson Street,
Southend-On-Sea, Essex, SS1 1EG
VAT Number: 848 7357 75
MOD FATS/II: FATS/2/MFL
DGFM Supplier Code: 46945