W3C home > Mailing lists > Public > public-hcls-coi@w3.org > January to March 2008

RE: An argument for bridging information models and ontologies at the syntactic level

From: Kashyap, Vipul <VKASHYAP1@PARTNERS.ORG>
Date: Thu, 27 Mar 2008 06:58:42 -0400
Message-ID: <DBA3C02EAD0DC14BBB667C345EE2D1240236D16C@PHSXMB20.partners.org>
To: "Samson Tu" <swt@stanford.edu>
Cc: "Ogbuji, Chimezie" <OGBUJIC@ccf.org>, <public-hcls-coi@w3.org>, <public-semweb-lifesci@w3.org>
	

It seems to me that we shouldn't throw out an important distinction just because
SNOMEDCT has terms that should be expressed in information model.
[VK] I did not mean to throw out the distinction, but to position the
distinction differently. For instance, I would still differentiate between
SNOMED-CT the information model
and SNOMED-CT the terminology, just that I would view SNOMED-CT the terminology
as a specialized case of the SNOMED-CT the information model. 
 
So, would propose that the notion of an information model "subsumes" the notion
of a terminology. 

To me, information model = structure and types of information (statements or
expressions uttered, recorded, written by somebody). 
[VK] This is part of the disagreement. How about:
Information Model = Semantics and types of information ...
 
 Informational entities (e.g., observations, eligibility criteria) make
reference to things in the world through terminology codes (e.g., 57054005 for
acute MI in SNOMEDCT). 
[VK] You don't need a terminology to do that. For example, Acute MI could be a
class in the Information Model. Further more it could be a sublcass of the
Observation/Problems class.
>From the modeling point of view it's not necessary, This decision of referencing
external terminology codes appears to be taken due to implementation and
scalability considerations..
 
 A good terminology system should have an ontology of things and relationships
in a particular domain. 
[VK] Agree
 
These things and relationship have labels (terminology codes) that are used by
entities in the information model. The codes are references to things in the
real world.
[VK] There is no reason (apart from implementation concerns) whey these things
and relationships can't be modeled in the information model and refer to things
in the real world.


		An information model can also be part of an ontology, if your
ontology needs to model  informational entities syntactically.
		[VK] I am not sure why you want to model informational entities
syntactically in an information model? Shouldn't you model the semantics of
these entities?
		I think the label "information model" may be a misnomer. In
general syntax and structure of data is modeled in data and implementation
models? 
		
		You need to have an information model of data you can query if
you want to write eligibility criteria or abstraction definitions. 
		[VK] Isn't an information model a "model of information"?
Representing the semantics of information should enable specification of
eligibility criteria and definitions?
		In my mind, a model of data is a data model and models the
structure and organization (as opposed to the semantics) of the data.
		 
		 For example, in an ontology of clinical trials, eligibility
criteria are information-content entities that you model syntactically, because
knowledge representation languages (such as OWL) are not expressive enough. 
		[VK] OK, this is an expressivity issue. In the case where
certain languages are unable to model the semantics, you could go to a different
language (rules or model them syntactically, but that is a different issue and
at that
		point the model becomes a data or implementation model.
		 
		---Vipul. 


The information transmitted in this electronic communication is intended only
for the person or entity to whom it is addressed and may contain confidential
and/or privileged material. Any review, retransmission, dissemination or other
use of or taking of any action in reliance upon this information by persons or
entities other than the intended recipient is prohibited. If you received this
information in error, please contact the Compliance HelpLine at 800-856-1983 and
properly dispose of this information.
Received on Thursday, 27 March 2008 10:59:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 27 March 2008 10:59:25 GMT