Re: RM's Canonical database
From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Tue, 04 Jul 2006 03:39:39 GMT
Message-ID: <%plqg.5907$pu3.133892_at_ursa-nb00s0.nbnet.nb.ca>
>>It is one thing to say, "given the limitations of the dbms we are
>>using,
>>and other practical considerations, we have decided to implement
>>business rules in a middle tier, and require all application code that
>>issues updates to use that tier." It is a different thing to say
>>"business
>>rules shouldn't go in the dbms."
>>
>>The first one is situationally-dependent; it might be a good idea
>>or not depending on a variety of practical considerations. The
>>second one is just false.
Date: Tue, 04 Jul 2006 03:39:39 GMT
Message-ID: <%plqg.5907$pu3.133892_at_ursa-nb00s0.nbnet.nb.ca>
Ron Jeffries wrote:
> On 3 Jul 2006 09:04:21 -0700, "Marshall" <marshall.spight_at_gmail.com> wrote: >
>>It is one thing to say, "given the limitations of the dbms we are
>>using,
>>and other practical considerations, we have decided to implement
>>business rules in a middle tier, and require all application code that
>>issues updates to use that tier." It is a different thing to say
>>"business
>>rules shouldn't go in the dbms."
>>
>>The first one is situationally-dependent; it might be a good idea
>>or not depending on a variety of practical considerations. The
>>second one is just false.
> > Repeatedly asserting that business rules should always go in the DBMS doesn't > make it more true. > > I'm not saying that they shouldn't ... but that /sometimes/ they shouldn't.
I suggest, when one considers what constitutes /sometimes/, it first requires a failure of the dbms in some manner. Received on Tue Jul 04 2006 - 05:39:39 CEST