Jim,
I fully understand the perspective in which you present this paper, and
agree with most of it. However, for those with a different background
it would seem to ignore many successful approaches to interoperability -
common data models being only one. We should better explain where these
approaches succeed and fail. For example; (01)
* Basic network connectivity (Hey, we just got this a few years ago!)
* File transfer and conversion programs (still many of these around)
* Distributed object and service based approaches based on ad-hoc
interfaces
* Standard interfaces/services for domain specific needs
* Data Warehouses
* Event brokers
* Agents
* Model driven Architecture based approaches
* Various mapping technologies (02)
Since every one of these approaches has talked about solving the same
problems, and has added some element of improvement, we should cover
what is and is not solved. (03)
On the "edge" there are the semantic web services approaches. (04)
There are a few key differentiators;
* Flexibility for adaptation in technology, structure or term
-- What if I just need to add elements or don't need some elements?
-- What if I have the same elements arranged differently
-- What if I use another term for the same element or the same term for
different elements?
* Adaptation across an architected solution Vs. adaptation between
architected solutions
-- By architected solution I include all forms of data, interface,
process, term and service specification - technology specific or not.
* Dynamic adaptation Vs. Manual or partially automated adaptation. (05)
You correctly identify 2 basic approaches;
* Deep description of concepts - the machine "understands" that things
are the same or related
* Shared concepts - the machine records the agreement that things are
the same or related (06)
I share your conclusion that shared concepts are the only practical
mechanism at this time (Some people I respect think differently). Upper
ontologies are one mechanism for shared concepts, as are standard
interfaces or common data models. Standard interfaces and data models
suffer from structural and term based fragility, where ontologies may be
less subject to these problems if used correctly. (07)
>From a problem-statement point of view I would also be clear on your
goal of fully automated Vs. "human in the loop" adaptation. From my
viewpoint the easier goal of very fast and easy, human in the loop
adaptation is much more tractable and has a huge payoff in terms of
current needs. To go to fully trusted and fully automated adaptations
between fully disjoint domains requires a degree of precision that will
be very hard to obtain in our tooling as well as in the domain
architectures we have today. (08)
So then I would state the incremental problem as one of providing for
quick and accurate automated assistance in adapting services or data
structures that may have different structures or terms but are otherwise
compatible in intent. We should also talk about integration where the
intent is similar but not the same, which is very common. It would be
usefull to then give some examples of where this additional capability
would be applicable. (09)
I would also propose that some form of very flexible system for managing
and linking to a shared concept set is fundamental to these goals.
Ontology languages and upper ontologies are candidates for that shared
concept system. In addition to the technologies there are interesting
problems in terms of the social and political aspects of maintaining
that shared concept base and the relationship between shared concepts in
different communities. (010)
-Cory Casanave (011)
-----Original Message-----
From: cuo-wg-bounces@xxxxxxxxxxxxxx
[mailto:cuo-wg-bounces@xxxxxxxxxxxxxx] On Behalf Of Schoening, James R
C-E LCMC CIO/G6
Sent: Tuesday, December 19, 2006 7:49 AM
To: 'common upper ontology working group'
Subject: [cuo-wg] How would you rate these semantic technologies?
ConferenceCall Wed Dec 20, 11:30-12:30 (012)
CDSI, (013)
The attached paper rates the following semantic technologies: (014)
Technology: Score: (See Appendix 2 for scale)
XML/Metadata 9
RDF 9
OWL 9
Stand-alone ontologies 6
Upper Ontologies 4
Mapped Upper Ontologies 1 (015)
What's your opinion? Join us for tomorrow's teleconference and
let's discuss. Details below (016)
Jim Schoening (017)
-----Original Message-----
From: Schoening, James R C-E LCMC CIO/G6
Sent: Thursday, December 14, 2006 8:45 AM
To: common upper ontology working group
Subject: CDSI Draft Paper and Conference Call Wed Dec 20, 11:30-12:30 (018)
Cross Domain Semantic Interoperability WG, (019)
The 2nd conference call of this group is scheduled for next
Wednesday, Dec 20th, 11:30 AM -12:30 PM EST. Join by dialing
(888)622-5357 and entering the code 342803. (020)
Attached is a draft paper, which will be discussed. The paper
attempts to explain, to the layman, that current technology cannot
achieve 'Data Interoperability across the Enterprise.' It concludes
that emerging technology has promise and should be pursued. It makes a
first cut at rating various technologies on a 9-level maturity scale,
which needs further discussion. (021)
Please join the telecon and invite other stakeholders. The
CDSI-WG web site is at http://www.visualknowledge.com/wiki/CDSI. (022)
James R. Schoening
U.S. Army C-E LCMC CIO/G6 Office
Voice: DSN 992-5812 or (732) 532-5812
Fax: DSN 992-7551 or (732) 532-7551
Email: James.Schoening@xxxxxxxxxxx (023)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/cuo-wg/
Subscribe/Unsubscribe/Config: http://colab.cim3.net/mailman/listinfo/cuo-wg/
To Post: mailto:cuo-wg@xxxxxxxxxxxxxx
Community Portal: http://colab.cim3.net/
Shared Files: http://colab.cim3.net/file/work/SICoP/cuo-wg/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/CommonUpperOntologyWG (024)
|