Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> comp.databases.theory -> Re: The fable of DEMETRIUS, CONSTRAINTICUS, and AUTOMATICUS

Re: The fable of DEMETRIUS, CONSTRAINTICUS, and AUTOMATICUS

From: Tony Andrews <andrewst_at_onetel.com>
Date: 19 Oct 2004 05:57:10 -0700
Message-ID: <1098190630.425442.147330@z14g2000cwz.googlegroups.com>


Kenneth Downs wrote:
> You want active automation anywhere were some derivation is taking
place.
> If you store extended price in your detai lines, a constraint saying
it
> must be price * qty is silly. The constraint should be that the user
may
> not directly write that column, and Automaticus fills it in for you.
>
> The thesis I am developing is that there are not that many places
where you
> should use a passive constraint, it may be possible to limit advanced
> complex systems solely to Unique and RI. The real service and power
of the
> system will be in the automatically generated information. If I am
wrong,
> I hope I've the gumption to report here why it did not work. If I'm
right,
> I'll be releasing something under GPL someday. Wish me luck :)

I do. What are are saying seems uncontroversial: that which the system can derive, it should derive; that which it cannot, the user must supply correctly. But there is a lot more than RI and uniqueness that cannot be derived: rules like "person must be aged 18 or over on start date", "person cannot be assigned to 2 or more departments at the same time", etc. etc. It isn't clear to me what new insight you have here that will lead to systems that can automatically pevent such rule violations by actively supplying the correct values. Received on Tue Oct 19 2004 - 07:57:10 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US