Bill, (01)
I too apologize for extending this out, but I think there are still
important issues here. After this, I suggest we take this offline. (02)
These two folks may indeed have understood the issues, but in general I
would say that data modelers don't. At least in my experience. I
remember advocating ontologies in 1988 at a company I worked for which
built a natural language interface to relational databases and targeted
companies in the insurance, pharmaceutical, etc., industries. I saw so
many contorted data models, schemas, and databases, all of which, say
in the insurance industry, could have been very close semantically and
dealt with a very similar real world, but weren't and didn't: the
databases were baroque, and we spent much time trying to figure out
their semantics, developing a mapping to 1) our homegrown database
expert system/knowledge base, from which 2) the natural language system
could interact. Back in the way old days I was a COBOL
programmer/analyst at a utilities company and the data models we had
then had many of the deformations I spoke of: they weren't just about
the world, they had system and data constructs interwoven. (03)
I would say that (and have in this thread) that conceptual
models/schemas are the most semantics you typically have when you build
a database. But using normal methodology, i.e., the 3 schema model and
normalization, you throw that semantics away in reducing to the
physical model. The only place you have left for the semantics is the
data dictionary, written in English, or in the code of surrounding
procedural programs, or in the very limited physical schema. Or in a
picture that you saved of the conceptual model/schema, but which has no
direct connection to the actual database anymore. (04)
Similarly, in the OO community, you develop a conceptual object model
that you then give to programmers to help them program their code, and
then you throw away the model, never maintain it, or keep it forever as
a picture. (05)
I too recall a major technical thread in the database community in the
late 80s and early 90s, on developing more semantic data models. I
know, because I read those papers then and I worked with data models
and databases. For example: (06)
J. Peckham and F. Maryanski. Semantic Data Models. ACM Computing
Surveys, 20(3):153--189, September 1988. (07)
http://portal.acm.org/citation.cfm?id=62062
From: The ACM Digital Library The Guide (08)
Semantic data models
Full text Pdf (3.10 MB)
Source ACM Computing Surveys (CSUR) archive
Volume 20 , Issue 3 (September 1988) table of content
Pages: 153 - 189
Year of Publication: 1988
ISSN:0360-0300
Authors Joan Peckham Univ. of Connecticut, Storrs
Fred Maryanski Univ. of Connecticut, Storrs
Publisher ACM Press New York, NY, USA (09)
ABSTRACT
Semantic data models have emerged from a requirement for more
expressive conceptual data models. Current generation data models lack
direct support for relationships, data abstraction, inheritance,
constraints, unstructured objects, and the dynamic properties of an
application. Although the need for data models with richer semantics is
widely recognized, no single approach has won general acceptance. This
paper describes the generic properties of semantic data models and
presents a representative selection of models that have been proposed
since the mid-1970s. In addition to explaining the features of the
individual models, guidelines are offered for the comparison of models.
The paper concludes with a discussion of future directions in the area
of conceptual data modeling. (010)
Object-Oriented databases were coming out about the same time. Why,
because more expressivity was needed. And by the way, the predominant
conceptual modelling language back then was ER, much as it is today,
though UML is finally beginning to crowd it out. (011)
It's also my experience that data modelers are one of the greatest
impeders of the adoption of semantic models and ontologies. They
already think they're doing that, so why introduce anything different?
But they're not doing that. Ask Barry about his experiences with
bio/medical databases. (012)
Thanks,
Leo (013)
_____________________________________________
Dr. Leo Obrst The MITRE Corporation, Information Semantics
lobrst@xxxxxxxxx Center for Innovative Computing & Informatics
Voice: 703-983-6770 7515 Colshire Drive, M/S H305
Fax: 703-983-1379 McLean, VA 22102-7508, USA (014)
-----Original Message-----
From: uos-convene-bounces@xxxxxxxxxxxxxxxx
[mailto:uos-convene-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Bill
Andersen
Sent: Monday, February 27, 2006 2:20 PM
To: Upper Ontology Summit convention
Subject: Re: [uos-convene] Other Approaches Too. (015)
Leo, (016)
At the risk of dragging this out... (017)
On Feb 27, 2006, at 13:39 , Obrst, Leo J. wrote: (018)
> LEO: Yes, I tried to make a rhetorically strong point by using
> 99% ;) I
> would say that data modelers start out trying to capture their local
> semantics in a conceptual model (I used this as synonymous with
> "conceptual schema" because that's my usage and most database folks).
> This is not necessarily about the real world, just their corner of
it,
> with entities such as purchasingDept, payroll, VSAM storage, employee
> and employee identifier, specific Bureau of Labor reports, etc. Yes,
> these items are all in the real world, but do they involve
ontological
> entities in the way we usually understand them? I think not. (019)
That's just false, Leo. I'd point you again to Bill Kent's "Data and
Reality" or David Hay's "Data Model Patterns: Conventions of
Thought". These guys seem to have a much better grip on ontology in
their recommendations for data modeling than that exhibited in 99% of
the "real" ontologies I've seen written in places like (020)
http://www.schemaweb.info/ (021)
> They do
> try to capture the local semantics needed for the local data
> store/application. (022)
Sometimes yes. Sometimes no. (023)
Bill Andersen (andersen@xxxxxxxxxxxxxxxxx)
Chief Scientist
Ontology Works, Inc. (www.ontologyworks.com)
3600 O'Donnell Street, Suite 600
Baltimore, MD 21224
Office: 410-675-1201
Cell: 443-858-6444 (024)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files:
http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit (025)
|