Re: SUO: questions about SUMO
Dear Matthew,
I hope I won't bother you too much, but your comments to Michael and Ian
are yet another *perfect* example of the positions I am challenging.
Quoting from: http://suo.ieee.org/email/msg08409.html
> OK. I think I should point out that time is not a real number, as
> Michael seems to be suggesting. Time is something that can be
> mapped to real number.
>
> That time can be usefully mapped to real number is a consequence
> of time and reals sharing some properties. In particular that all
> points in time are part of a continuum (at least this is how this
> mapping represents time). (This is true for all properties that can
> be expressed by a number in a unit of measure. The unit of measure
> is the name of the mapping between the property and the number
> space).
>
> In fact there are multiple possible mappings of time to real number,
> depending on the unit of time that is used, and almost certainly it
> is to a subset of real numbers. (Current knowledge suggests that the
> universe came into existence some 13b years ago, in which case the
> lower limit will at least not be minus infinity).
I DO agree about most of what you say above, but my position is
that this is NOT what should be done to advance toward the solution
of ontologies interoperability:
1) You are arguing about the "meaning" of the ontology content,
not the ontology structure. This is a perfectly well debatable
question, and THIS kind of questions WILL have to be answered,
some day... by the USERS of "temporal representation" ontologies.
2) It is NOT your "job" nor the job of any SUO WG or like group to
settle such questions. You may want to choose answers about those
touchy questions of "meaning" for *your* ontologies, the ones you
use or intend to use in your own personal turf. But you are not
entitled (nor anyone else for that matter!) to enforce your choices to
other parties, nor are you likely to succeed with such an endeavour.
3) These considerations about "meaning" have nearly *no* relevance to
the design of ontologies interoperability. I will try to explain why
below, using a metaphor.
A "plumber" does not care that much about what flows thru the pipes he
is installing, it could be water, oil, fancy chemicals, steam, whatever.
Of course the "nature" of the carried fluid *does* matter a bit, but,
somehow *just a bit*. The ducts have to stand the pressure, viscosity,
and corrosiveness of the fluid, but HOW MUCH of the technical specs
does this account for with respect to overall "ducts engineering"?
This is exactly analogous for ontologies!
Yes, someone will have to choose about the "supported" meanings, and,
*some* considerations should be given to what those meanings could be.
But these meanings DON'T have to be FIXED to proceed to ontologies
interoperability. We just need to have a "fair" idea of what they
will "look" like (just the pressure, viscosity and corrosiveness)
in order to design the interoperability capabilities.
Not only these characteristics of the ontologies contents brings only
a *very little* part of the design specs, but refining on the very
precise meanings within the anticipated ontologies brings NOTHING.
Just like the "plumbing" can withstand any other kind of fluid which
is "close enough" to the one it was designed for.
This is the reason why, as well as I believe that IFF *could* be an
usefull starting point for ontologies interoperability (if used as a
kind of a "meta framework"), I firmly maintain that "merged"
ontologies like SUMO are TOTALLY IRRELEVANT.
Some people are boneheaded, but I will be mercifull and hope
for them that the bone is harder than the bricks of the wall!
Cheers.
-- Jean-Luc Delatre
--------------------------------------------------------------------------
"The sage (the person in accordance with Tao) have no particular interest.
They see all people (equally) as straw dogs." - Lao Tzu: Tao Te Ching
From N. 5 in http://www.csudh.edu/phenom_studies/laotzu/taoteching.htm
--------------------------------------------------------------------------
http://perso.club-internet.fr/jld/ -- GSM: +33 6 11 24 06 29