Re: Warning about null and open question to Oracle
Date: 1995/06/18
Message-ID: <3s0k3f$egq_at_usenet.interramp.com>#1/1
In article <3rnui6$mdf_at_News1.mcs.com>, stowe_at_mcs.net says...
>
>
>... Since a NULL is not TRUE, it will not behave like a TRUE result. Some
> people seem to forget this, or never learn it in the first place.
>
I agree. After 2 dozen responses over 5 days this thread has reached no
consensus...for good reason. The general views seem to be:
- "It seems to me, nulls should work like...come on Oracle"
- "Sorry, Oracle is SQL standard compliant with respect to nulls..."
- "Forget nulls..."
It would seem to me that you are all very interested in getting the job done. My advice to those interested in trying to learn nulls: Master the relational theory. To those who think they've got a better idea: Maybe. But unless you know the 25 year history of relational theory, you probably are suggesting nothing new. To those not inclined to study the theory: You're armchair quarterbacks. Get serious or get out of the business. You're doing your users a disservice. To those who defend three-valued logic (that's nulls for you quarterbacks): Good luck trying to get a user to take a course on set theory. If they don't understand nulls properly (and a quick glance at this thread suggests they won't), there's a good chance they will get wrong answers. Users need to get serious too, but three and four-value logic is a bit of a stretch.
Tom Jamieson
pp003009_at_interramp.com
Received on Sun Jun 18 1995 - 00:00:00 CEST