Re: OO and relation "impedance mismatch"
From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Tue, 05 Oct 2004 12:49:47 -0400
Message-ID: <cbjujc.dg3.ln_at_mercury.downsfam.net>
>
> Wouldn't it be at least theoretically possible to get all the errors but
> still work declaratively? Maybe the difference is simply asking the
> DBMS to report all declarative violations instead of just the first
> one.
>
> I am speaking hypothetically, not about any specific product.
Date: Tue, 05 Oct 2004 12:49:47 -0400
Message-ID: <cbjujc.dg3.ln_at_mercury.downsfam.net>
Marshall Spight wrote:
> "Kenneth Downs" <firstinit.lastname_at_lastnameplusfam.net> wrote in message
> news:ohjsjc.99t.ln_at_mercury.downsfam.net...
>> >> Ironically, this led me to ask how serious a performance penalty I would >> get if I discarded declarative RI and went back to the old-fashioned >> method of using triggers.
>
> Wouldn't it be at least theoretically possible to get all the errors but
> still work declaratively? Maybe the difference is simply asking the
> DBMS to report all declarative violations instead of just the first
> one.
>
> I am speaking hypothetically, not about any specific product.
Sure, seems it would be an implementation decision for the DBMS vendor, and would be orthogonal to decisions about how to implement the RDM. Alas for me, the world has gone fail-on-first in all products I've seen.
Well, the only exception would be compilers, which tend to a tell-me-everything approach, running through the entire program and reporting all errors without stopping.
>
>
> Marshall
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Tue Oct 05 2004 - 18:49:47 CEST