Dear Cory, (01)
With respect to your 3rd point, "confusing people and roles", please
find attached a recent paper of mine at the last FLAIRS (May 2011).
Regards, (02)
G.P. Zarri (03)
Cory Casanave a écrit :
> Jim,
> I was also involved with the OMG "person" effort Ed is referencing and have a
>somewhat different perspective (which is not unusual for Ed & ). My
>perspective is that there is a quantifiable set of reasons why this and
>similar efforts failed:
>
> 1: A structural approach
> The group of people were used to defining CORBA interfaces or, perhaps,
>Object models. The problem they ran into is that they could not "agree" on
>any particular set of properties of a person. This is, in part, due to an
>expectation that the definitions of person defined a data structure rather
>than a set of possible properties that could be used, as required, in any data
>structure for any purpose.
>
> 2: Labels rather than concepts
> Due to the same technology orientation there was a lot of concern about the
>"one true" name for things rather than accepting that there could be multiple
>names for the same concept.
>
> 3: Confusing people and roles
> When stakeholders start talking about people they are frequently talking
>about a role that person may play. Trying to join every possible role of a
>person into a single element becomes intractable and complex. Separating role
>types from the physical person clears up many of these issues. It also solves
>a lot of the "agreement" issues since roles come with some degree of context.
>
> 4: Data types not Units
> As required by most of our technologies the types of properties are
>computational types such as "int", "Real" or "String". Had they focused on
>the quantities and units required to represent properties of people they would
>have been able to encompass the different representations of numbers, dates,
>locations, etc.
>
> Had the group focused on the concepts relating to persons and the roles they
>play rather than static structures with fixed names it is my belief that a
>useful outcome would have been possible. They would then have been able to
>relate those concepts to the various data structures that represent people and
>the roles they play. I am sure many on this list recognize these as
>differentials between conceptual and data modeling.
>
> I would be interested in participating in your effort. I am also
>participating in an OMG effort called "Semantic Information Modeling for
>Federation" (SIMF) to provide standards for conceptual and logical models with
>the required bridging relations between those models. These efforts may be
>collaborative.
>
> -Cory Casanave
>
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Ed Barkmeyer
> Sent: Tuesday, July 12, 2011 10:34 AM
> To: [ontolog-forum]
> Cc: Larry L. Johnson; Chuck Allen
> Subject: Re: [ontolog-forum] Person Ontology Project
>
> Jim,
>
> I wish you good fortune, but I think you need a very careful constraint on
>the scope and purpose of the 'person ontology' to have any hope of success.
>
> About 10 years ago, there was active standardization of many domain-specific
>models in the Object Management Group -- financial objects, workflows,
>"product data management" (authorizations, assignments), medical information
>access, telecomm business info, etc.
> A friend and longtime standards professional, Larry Johnson, organized a
>"People Who Like People" get together at each quarterly meeting, with the
>purpose of getting all the views in front of all the viewers, and hopefully
>coming to some common base model. After something like 8 meetings, we
>abandoned the effort. It takes a lot of work to establish a viable common
>base model that supports all the viewpoint models.
>
> Similarly, after 7 years of work and 20+ standards, the HR-XML consortium
>(Human Resources information sets) decided that for version 3 of the HR-XML
>standards suite, it was important to create a common base model, because most
>of the principal concepts had different terms and somewhat different
>interpretations in the separate HR-XML standards (because they were produced
>by different working groups at different times). To the best of my knowledge,
>they are 2+ years into the effort and still struggling.
>
> And these are groups that are/were modeling at the UML level. If you want to
>put formal axioms into your ontology, it becomes even more difficult to get
>agreement. There were only 6 views of the legendary elephant; there are many
>more views of 'person'.
>
>
> Best of luck!
>
> -Ed
>
> (04)
GPZ Flairs 2011.pdf
Description: Adobe PDF document
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
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 (01)
|