Re: OO and relation "impedance mismatch"
Date: Tue, 05 Oct 2004 09:47:03 -0400
Message-ID: <nk8ujc.jh2.ln_at_mercury.downsfam.net>
Troels Arvin wrote:
> On Tue, 05 Oct 2004 08:45:00 -0400, Kenneth Downs wrote:
>
>> Here is an example of my current experimental form.
> [...]
>
> INSERT statements certainly become less simple this way :-)
? Maybe I don't know what you mean, but the specific design principle is that the person accessing the database with insert/update commands does not need to know or do anything special. So I can code an app that uses .Net datasets, or I can go into SQL Query Analyzer and type "Insert into ... ".
The only thing I need to know when I perform an Insert is what values to put into what columns.
> - But it looks like most of the details can be delegated to a wrapper
> function in the application (or perhaps in the DBMS?).
I fall into that camp that likes the biz rules inside the DBMS server. Otherwise interactive tools become a back door that allows people to put bad data into the database. It is also ultimately more efficient, but that claim can spawn an entire thread on its own.
>
> Thanks; very interesting.
>
anytime.
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Tue Oct 05 2004 - 15:47:03 CEST