Re: [ontac-forum] Re: Semantic Layers (Was Interpretation of RDF reification)

Henry,

i don't ignore you messages, moreover, when i can try to read them 
attentively; for they always have some useful learning about practical 
issues.
See some comments below.

Respects,
Azamat Abdoullaev

----- Original Message ----- 
From: "Henry Story" <henry.story@bblfish.net>
To: "Azamat" <abdoul@cytanet.com.cy>
Cc: "Harry Halpin" <hhalpin@ibiblio.org>; <semantic-web@w3.org>
Sent: Tuesday, April 04, 2006 12:11 PM
Subject: Re: [ontac-forum] Re: Semantic Layers (Was Interpretation of RDF 
reification)



Hi Azamat, since you are in the mood of giving lessons, perhaps you
should also not ignore replies made to you.

I posted a reply to you on this thread, and you completely ignored
it. I also want my 5 page reply ;-)

There are 3 points:

a. It is possible to pass from very contextual data to less
contextual data [1]
b. It may be impossible to completely decontextualise data [2]
c. In another thread someone seems to in fact be embarked on
creating a very decontextualised ontology
   using RDF of the type you claim is impossible to do with these
tools [3]

Because of (a) above it is ok for people to encode their data on the
semantic web in a very contextual fashion. Ie. Nothing stops people
just getting their work done. Furthermore there is no need to wait
for the final upper ontology to be complete, since it may be
impossible.

Here may lie all the misreading of my point, which i would like to 
emphasize. There may be a multitude of different ontologies, perspectives, 
views, pictures, attitudes, outlooks regarding the nature of 'environmental 
universals' as substance, object, state, change, time, space, cause, 
quality, relation, body, etc. Still among human beings and human races, 
there is an established agreement regarding some general things in the 
universe; namely:

The world or reality or universe exists;
There are entities or things or beings in the world;
Entities have properties;
Entities have parts;
Entities stand in relationships;
There are space-time relations;
Things change over time;
Entities exist as substances, states, and processes
Substances have masses and occupy spaces; etc.

The collection of these common ontological axioms make the general knowledge 
about reality, or the ontological context, the general commitments 
underlying all sorts of specific contexts, models, theories, and languages, 
like SW formal languages. Nobody is here talking about a single ontological 
standard dictating how you should express your conceptual world, it may be 
as chaotic as you like. But when you model some real world domain or 
communicate with somebody or try to understand some natural language, the 
meanings of its words. Above your personal commitments, you need to evoke 
(represent or code) a general knowledge about things [the result of your 
perception, deep learning, experience, and reasoning] unifying all reasoning 
entities, human and machine, within a single intelligent framework. This is 
the primary meaning of UFO, unifying framework ontology of the world and its 
content, which is designed to bring some fundamental order, organization and 
structure to your (Web) data, and so ontologically contextualized.


On the other hand nothing stops anyone working on a very
decontextualised ontology, which when finished, everyone will be able
to map their data to using [1] if necessary. So stuff like (c) is
welcome and it seems indeed possible.

So you are left with trying to show that the work done in (c) is
failing in some essential way. And you have to proove that this
failing is such that no more decontextualised framework can be
created that will help resolve this problem. Then you would also have
to show that no useful work can be done at all even without such a
generalised schema.
For me to feel that you had prooved this, I would have to believe
that you had understood rdf very well. But at present you seem to be
stumbling on some basic misconceptions such that one cannot express
nary relations in rdf,

See my distinction of ontological relations and formal logical relations, 
which is only a variety of relationships [sent to Hans].


even though I pointed this out recently [4]

[[
If you want to relate an object to a bunch of things just  relate it
to a list.

:henry :onlyowns (:laptop :hat :camera) .

or relate it to a class with a number of properties

[ :color :blue;
    :temperature "25";
    :position [ :lat "23"; :long "45" ]
] .

]]


Henry Story


[1] See code http://lists.w3.org/Archives/Public/semantic-web/2006Mar/
0175.html
[2] mentioned in this paper http://dig.csail.mit.edu/breadcrumbs/node/
101
     but probably the result of Gödel's incompleteness theorem. Good
arguments for this somewhere anyone?
[3] http://lists.w3.org/Archives/Public/semantic-web/2006Mar/0312
[4] http://lists.w3.org/Archives/Public/semantic-web/2006Mar/0252

Received on Tuesday, 4 April 2006 14:57:22 UTC