RE: SUO: a silly question about the new modular architecture
Dear John,
I'll try to give how I see this.
Matthew West
Principal Consultant
Shell Information Technology International Limited
Shell Centre, London SE1 7NA, United Kingdom
Tel: +44 20 7934 4490 Other Tel: +44 7796 336538
Email: matthew.r.west@is.shell.com
Internet: http://www.shell.com
> -----Original Message-----
> From: Thompson, John A [mailto:john.a.thompson@Boeing.com]
> Sent: 07 September 2001 18:26
> To: SUO
> Subject: RE: SUO: a silly question about the new modular architecture
>
>
>
> I'd like to explore the ripple effects of having both a 3d
> and a 4d module a little bit more.
>
> John Sowa replied (in part):
> -------------------
> The modular approach would minimize the amount of overlap, so there
> would actually be very little recoding that would need to be done.
> For example, all the axioms about geometry, space, time, etc.,
> could be used unchanged. Axioms for mathematical structures,
> physics, engineering, etc., would be unchanged if they used the
> 4D coordinate system.
> -------------------
MW: Note the last phrase. Maximising commonality will not just drop
out, it will need working at. This will improve overall quality though.
>
> This seems to be saying that we'd need only one version of
> the theories for geometry, space, time, mathematical
> structures, physics, and engineering, and that they would
> work equally well with either a 3d or 4d module. What is it
> about these theories that makes them immune to a change
> between 3d and 4d?
MW: I'm not sure all of these would be imune, but ones that would be
imune would be ones about classes/sets. A set does not change its
nature because of 3D and 4D.
> What other theories would not be immune
> and would need two versions?
MW: 3D and 4D are essentially different theories about individuals,
although activities/processes turns out to be at least very similar.
So I suspect that physics and engineering would not be imune.
> My intuition is that most of
> the ontology modules would need two versions, as they would
> be based on either on the 3d or 4d module as a parent theory.
MW: I've recently been converting from 3D to 4D data models, and my
estimate is that two thirds to three quarters is different.
>
> What exactly would be in the 3d and the 4d base modules? Can
> someone sketch out those two theories for us?
MW: Well actually Adam's recent notes were quite a good informal
statement of the sorts of things you would expect.
MW: A key part of a 4D ontology would be a temporal mereology. This
would include some interesting things like temporal parts inheriting
properties from those things they were temporal parts of.
>
> I assume that the lattice of theories would include many
> other basic choices in addition to the 3d vs. 4d choice. Can
> we list some of other basic choices that we'll encounter?
MW: Well Pat pointed out that the choice of identity basis was
strictly orthogonal to the 3D/4D choice (however in practice there
is a strong correlation with the choices people make with 3D/4D.
MW: But apart from that I don't think there is any choice as big.
At a lower level you will find things like CSG and BREP geometry,
and again the right thing will be to include both, and let people
choose what they want to use when.
> Won't the combination of all these choices require many
> versions of each lower theory that is based on them?
MW: Theoretically, but I don't think there are many choices, and
we only have to explore those that people are interested in. However,
it would be polite to construct things in such a way that others
could add what they thought was important later.
>
> I like the idea of the modular theories, but I'm concerned
> (like Adam) about the ripple effects of multiple choices for
> base theories.
MW: What I notice about reality is that if you try to ignore it
it eventually bites you. Better to take account of it but perhaps
not insist on doing everything immediately.
>
> Thanks,
> John Thompson
> Boeing
>