Message 12661 of the SUO list

Subject: Re: Status of SUMO?
Date: Thu, 24 Jun 2004 18:09:09 +1000
From: Philippe Martin
In-reply-to: John A. Thompson (private e-mail) and msg12632 by Adam Pease



Thompson to Adam:
> Sorry, but I do believe that names are fairly important in gaining
> broad acceptance of an ontology.

I have accepted Adam's reason for not wanting to change the name or the order
of the parameters of the "instance" and "subclass" relations in the SUMO: because
some communities like it that way.
However, because I prefer the convention used in graph-based languages, frame-based
languages or controlled English, and because in the MSO I integrate other ontologies
that have relations similar to those in the SUMO, you will have no problem to find
many lexical and semantic alternatives to the relations of SUMO. For example you can
use pm#kind or rdf#type instead of the more general sumo#instance (in my opinion,
sumo#instance is over-general for knowledge modelling but it may be useful for other
uses related to knowledge sharing). Below is an extract related to instance
and subclass relations (without their informal annotations). For more details and
other "relations from types", see
http://www.webkb.org/bin/categSearch.cgi?categ=pm%23relation_from_type+&recursLink=%3E&depth=4&hyperlinks
Interconnecting those relations coming from different ontologies is tricky since
they have different arguments and slightly different semantics. I may have got
it wrong. If you spot an error, please tell me.

Philippe


dc#Type__type (?, pm#entity)   //"dc#Type" is the identifier
  > sumo#instance;             //"Type" and "type" are just names

    sumo#instance (?, sumo#set_or_class)
     > pm#kind (pm)  sumo#immediate_instance;

       pm#kind__type__class__instance_of (?, rdfs#class)
        - pm#instance,
        = rdf#type  daml#type;

pm#subclass_of_or_equal (pm#set_or_class, pm#set_or_class )
 > sumo#subclass  rdfs#sub_class_of;

   sumo#subclass (sumo#set_or_class, sumo#set_or_class )
    > sumo#immediate_subclass;

pm#subtype__subtype_or_equal (pm#type, pm#type)
 > pm#strict_subtype  dl#subsumes_leaf,
 - pm#supertype,
 = dl#subsumes__SB;

pm#supertype (pm#type, pm#type)
 > rdfs#sub_class_of  sumo#subrelation,
 - pm#subtype;

   sumo#subrelation (pm#relation_type, pm#relation_type)
     > rdfs#sub_property_of,
     = cyc#genl_preds;