[Top] [All Lists]

Re: [ontolog-test] [ontolog-forum] is-part-of: a really, really, bad pra

To: <ontolog-test@xxxxxxxxxxxxxxxx>
From: <jmcclure@xxxxxxxxxxxxxx>
Date: Tue, 21 May 2013 08:30:13 -0800
Message-id: <00a401ce5640$792b2750$0200a8c0@McClure>
Thank you for your reply, John. You ask "What benefits, if any, would be
gained by mapping that logic to RDF?"    (01)

I'm in the RDF/SMW world - these are my hammers. I'm only looking for
practical information I can use in this particular world; that's the only
reason I am here. Maybe there's a disconnect between this world and one that
with booleans and such you indicate is necessary to create a full logic. But
I'm not trying to create a full logic, I only wish to create simple
applications users can use, ones whose information can be exchanged with
others.    (02)

The problem I'm hoping to hear input about is that ontologies seem to be
creating an unnecessary set of entities which are ballooning their
complexity with definite impacts on cost useability and adoptability. That's
the problem. And I am proposing a specific solution, one that facilitates
identifying is-part-of relations while preserving the relative efficiencies
of constrained English.    (03)

To wit, is-part-of is but a name for the relationship from a Part to a
Whole, that is, from an entity that is-a Part to one that is-a Whole; there
is no compelling reason I know why this needs to be the name of the
predicate used to associate Part to Whole. Instead, my proposal is to define
is-part-of as one or more directed graph patterns that, when detected,
indicate Something-Deemed-a-Part is-part-of Something-Deemed-a-Whole. This
would allow simple relationship connectors such as prepositions to be used
in practice, with the additional benefit that now multiple DAG patterns can
be defined by which the relationship from Part to Whole can be detected.    (04)

How is this so wrong? My objective is to create simpler systems for most
people; the massive duplication of names of classes as names of properties,
is just so obviously wrong -- it forces people to learn foreign concepts
like is-part-of just to enter instance information. How much easier to say a
Part is within a Whole, in a Whole or of a Whole! It's a bit frustrating
nothing is being said about that central conclusion of my analysis of many
ontologies.    (05)

Occam's Razor has direct application here.
Thanks - jmc    (06)

>-----Original Message-----
>From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx]On Behalf Of John F Sowa
>Sent: Tuesday, May 21, 2013 6:01 AM
>To: ontolog-forum@xxxxxxxxxxxxxxxx
>Subject: Re: [ontolog-forum] is-part-of: a really, really, bad
>On 5/20/2013 7:43 PM, jmcclure@xxxxxxxxxxxxxx wrote:
>> *Tenseless Amodal Properties.* I'd like to know how you'd model
>> this in RDF triples then
>> past(Harry believes (John Know Bill))
>RDF was designed for storing simple kinds of data in a database
>that could be freely intermixed with other stuff in a web page.
>The subset of logic represented in RDF has only two operators:
>existence and conjunction.  Two or more triples in the same
>page have implicit conjunctions among them.
>A blank node is an implicit existential quantifier.  It says
>that something exists that meets the criteria specified in its
>triple (and the other triples in the same page).
>As RDFS shows, it is possible to use RDF at the metalevel
>to make certain kinds of logical statements.  The use of RDF
>at the metalevel creates options for more expressive power.
>But RDFS has the same logical limitations as RDF, and it
>cannot compensate for the fundamental limitations of RDF.
>> I know for me it'd be:
>> Person:Harry had [[Belief:That JohnKnowsBill]]
>> Belief:ThatJohnKnowsBill of [[Type:Belief]]
>> Belief:ThatJohnKnowsBill of [[Type:Clause]]
>> Belief:ThatJohnKnowsBill subject [[Person:John]]
>> Belief:ThatJohnKnowsBill predicate [[Property:foaf:knows]]
>> Belief:ThatJohnKnowsBill object [[Person:Bill]]
>This is basically a translation of the syntax of a particular
>English sentence (or the simple logic-like notation) to RDF.
>To be a logic, you also need systematic ways of representing
>the Boolean operators and quantifiers of that subset of logic,
>rules of inference that determine what inferences are permitted,
>and a model-theoretic semantics that can determine truth values
>and can support proofs of soundness for the rules of inference.
>If you want an epistemic logic (for knowledge and belief),
>you need all of the above plus a lot more machinery for
>representing whatever theory of knowledge you choose.
>Summary:  If you want to represent logic, start by studying
>logic in more general and more readable notations.  Then you
>can decide what, if any, of that logic can or should be
>mapped to a highly restricted syntax, such as RDF.
>Even more importantly, ask yourself "Why?"  What benefits,
>if any, would be gained by mapping that logic to RDF?
>Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
>Config Subscr:
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (07)

To  Post: mailto:ontolog-test@xxxxxxxxxxxxxxxx
Msg Archives: http://ontolog.cim3.net/forum/ontolog-test/
Community Wiki: http://ontolog.cim3.net/wiki/
Community Files: http://ontolog.cim3.net/file/
Community Portal: http://ontolog.cim3.net/    (08)
<Prev in Thread] Current Thread [Next in Thread>
  • Re: [ontolog-test] [ontolog-forum] is-part-of: a really, really, bad practice?, jmcclure <=