Re: Business objects, subset of collection

Responding to Frebe...

It is trivial to develop a schema tailored to a
specific application that will be unusable by another application
solving a different problem and needing the same data.

It is also trivial not to do.

That is only true if one does NOT customize the schema to a particular problem. That's why RDBs should be constructed around the problem domain rather than specific problems. Whenever the schema is customized to a particular problem, one can usually find another problem in the problem domain where the schema is not usable.

Do you have references to a definition of CRUD/USER processing? If we
don't have a working definition, we use of the term seem to be rather
pointless.
Google it. You'll find plenty references to CRUD and USER acronyms.
I did. But didn't find any definition that could be useful for telling
if an application is CRUD or not. But obviously you would not have any
problem providing such definition.
If the primary tasks of the application are the acronym activities, it
is CRUD/USER. Order entry for a POS system is CRUD/USER. If the acronym
activities are peripheral to the problem being solved, the application
is not CRUD/USER. The forecasting, transportation, and other functions
of an inventory control system are not CRUD/USER.

Is this CRUD?

select i.invoiceid
from invoice i join payment p on i.invoiceid=p.invoiceid
group by i.invoiceid
having sum(p.amount) < i.amount and datediff(now(), i,duedate) >= 10

If that's all the application is doing, yes.

How do you apply 2NF to your classes?
The same as you do. Every non-identity responsibility must be a simple
domain (1NF) and must be fully dependent on the object identity.
"each nonkey attribute in the relation must be functionally dependent
upon the primary key."
Object identity is not the same thing as a primary key.
OTOH, a primary key is a tuple identity. They are semantically the same
thing in the RDM context.

The primary key is part of the tuple. Object identity (RAM pointer) is
not.

Whether it is embedded in the tuple or not, it still represents unique identity for the tuple and that is that is what is semantically important to the RDM. Only when the RDM is applied to RDBs does one need embedded keys.

--
There is nothing wrong with me that could
not be cured by a capful of Drano.

H. S. Lahman
hsl@xxxxxxxxxxxxxxxxx
Pathfinder Solutions
http://www.pathfindermda.com
blog: http://pathfinderpeople.blogs.com/hslahman
"Model-Based Translation: The Next Step in Agile Development". Email
info@xxxxxxxxxxxxxxxxx for your copy.
Pathfinder is hiring: http://www.pathfindermda.com/about_us/careers_pos3.php.
(888)OOA-PATH
.

Relevant Pages