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

Re: Multi-layered Knowledge Representations for Healthcare (was RE: An argument for bridging information models and ontologies at the syntactic level)

From: Dan Russler <dan.russler@oracle.com>
Date: Thu, 24 Jul 2008 09:30:40 -0400
Message-ID: <48888400.9020107@oracle.com>
To: "Kashyap, Vipul" <VKASHYAP1@PARTNERS.ORG>
CC: Samson Tu <swt@stanford.edu>, "Elkin, Peter L., M.D." <Elkin.Peter@mayo.edu>, public-semweb-lifesci@w3.org, public-hcls-coi@w3.org
See below...Yes let's move on to more examples...Dan

Kashyap, Vipul wrote:

>  
>
>     I agree with you below, except I think it's peoples'
>     "interpretation" of the RIM that causes the confusion, e.g. "The
>     focus of the RIM classes had primarily been structure and not
>     semantics." Since RIM is communicated in UML, UML semantic rules
>     apply, and one needs to be strict on the UMLsemantics  in order to
>     interpret the RIM.
>     [VK] Also, one needs to be make sure that the semantics of the
>     various classes in the RIM, e.g., Observation are consistent with
>     the semantics of the UML constructs being used. Furthermore, these
>     semantics/definitions of the various classes should be properly
>     defined in the RIM specs.
>
<dan> agreed...It would be helpful to start from scratch (with Act) and 
then add the constraint of Acts of subtype Observation. />

>
>
>     One way to think of "structure" is to think of the "structure" of
>     a definition in a dictionary, i.e. meaning one, meaning two, etc.
>     Then think of UML as another kind of structure for definitions,
>     i.e. a class representing a concept sets up a definition structure
>     of attributes, associations, and state transitions that define the
>     concept (analogous to the linguistic structure of "meaning one" in
>     a dictionary).
>     [VK] I think this is a stretch. Clearly there is widely accepted
>     connotation of the word "structure". A better approach would be to
>     introduce the notion of semantics and differentiate it from
>     structure explicitly across the RIM Specs.
>
<dan> I'm very interested in the "widely accepted connotation of the 
word 'structure.'" How wide is the acceptance. What is the connotation 
of 'structure?" />

>
>
>     I think that revisiting the RIM definitions and RIM UML diagrams
>     and rethinking of them as complementary "definitions of concepts"
>     rather than introducing set theory and other data management tasks
>     into the UML interpretation would be helpful.
>     [VK] I agree with the above, except that set theory is a
>     very useful mathematical tool to concretely specify the semantics
>     of various artifacts and has been broadly used in Computer
>     Science. Also agree with separation of data management issues from
>     information modeling
>     issues.
>
<dan> set theory is useful when thinking about sets of  instances of 
concepts; Also, agreed that it is widely used, not only in branches of 
computer science (especially data processing), but in many branches of 
science in general.  However, you will have to illustrate how the idea 
of set theory helps in understanding the semantics of a class in UML. To 
get us started, I'll take a stab with an example where set theory might 
help:
In UML, I can create a class to represent "DogsInMyKennel." I can select 
a list of attributes that are common across all dogs in my kennel, e.g. 
pen location, gender, etc. I can then create a subtype (child) of 
"PurebredDogsInMyKennel, that adds (by inheritance) the attribute 
DogBreed (value set = list of purebred dogs) to the set of attributes of 
the parent class DogsInMyKennel. Set theory is valuable in the sense 
that, logically, the number of instances of PurebredDogs cannot be 
larger than the number of instances of Dogs. However, one can't tell 
from the model, whether the number of PurebredDogs is the same or 
smaller than the number of DogsInMyKennel. Therefore, we can say that a 
"subtype" of a parent class in UML does not correspond to a "subset" of 
a set of instances of the parent class in UML unless one includes the 
case that when Set SubType = Set Parent, Set SubType is also a proper 
subset of Set Parent. Of course, one normally doesn't create a 
parent-child relationship in UML if there is only one child to the 
parent. So in UML, we wouldn't normally include the case that Set 
SubType = Set Parent.

hope that helps get us started.....Dan/>

>      
>     Seems to me that we need to do this exercise at least for a few
>     use cases, before we proceed on our attempts at determining
>     semantic conformance.
>      
>     ---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, 24 July 2008 13:32:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 24 July 2008 13:32:26 GMT