<7mkuj9$slf- at egroups.com> wrote:
>> In the OO methodologies I'm familiar with, one strives to get the data
> model to at least Second Normal Form. It's against the rules to have
> some Printer attributes identified by a (Printer identifier,
> document-format) tuple, while other Printer attributes depend only on
> the Printer identifier. Failing to correct this leads to the kind of
> problems we see now with document-format.
P.S. This book (http://www1.fatbrain.com/asp/bookinfo/bookinfo.asp?the
Object-Oriented Systems Analysis : Modeling The World In Data
By Shlaer, Sally / Mellor, Stephen J .
Softcover; 144 Pages
Published by Prentice Hall
Date Published: 04/1989
contains an excellent treatment of this topic. Obviously, it's a
Shlaer/Mellor book, but the techniques descibed apply equally well to
other methodologies such as Rumbaugh OMT or UML. In fact, the
techniques originate in the database world.