Re: So what's null then if it's not nothing?

From: mountain man <hobbit_at_southern_seaweed.com.op>
Date: Wed, 23 Nov 2005 02:07:47 GMT
Message-ID: <T9Qgf.1119$ea6.357@news-server.bigpond.net.au>

"paul c" <toledobythesea_at_oohay.ac> wrote in message news:OhPgf.573740$1i.24270_at_pd7tw2no...

> Frank Hamersley wrote:

>> JOG wrote:
>>
>>> michael_at_preece.net wrote:
>>>
>>>> "A value of the Boolean data type is either true or false. The truth
>>>> value of unknown is sometimes represented by the null value."
>>>>
>>>> Dodgy or what?! It would be better to use a large cumbersome robot
>>>> waving its arms about saying "does not compute" than to use null for
>>>> this.
>>>>
>>>> Mike.
>>>
>>>
>>> I for one _really_ hope some database corporate representatives are
>>> reading this thread, and take this feature into consideration for their
>>> future releases.
>>
>>
>> Errr...pardon me for being dense...but exactly what is the "feature" we
>> so desperately need?
>>
>> Cheers, Frank.

>
> I was trying to exercise restraint, but I guess I've gotten sucked in 
> along with some smart people.  puts me in good company, i suppose ...
>
> i don't understand why such a question ends up with so much space on the 
> "databases.theory" newsgroup.  is it because there is no such group as 
> "comp.databases.practice.etc"?

Its because the theory side of databases is lead at the moment by a large hand waving robot called "The Pedagogy of the RM". This practice (as most human concerns go) is self perpetuating and has been at it in excess of 30 years.

The problem is that evolution in the world external to the academic vision of the relational model has not been matched by evolution of the model, or evolution of its interfaces to other valid models and ideas in the (information technology) environment.

Technically, the NULL can mean a different thing to every DBMS and its treatment (if any) will vary between vendors. Tabulate the vendors of all RDBMS and DBMS software and you will see that this is a fact.

Some people believe that unknown and unknowable values cannot exist, and they might be happy to contemplate nulless existence in a vendor software that represents their notion of reality.

Others believe that it is better to allow for the possibility of unknown and unknowable values, and select and/or write their RDBMS software accordingly.

None of this is against the law (yet), so opinion is thus diverse.

-- 
Pete Brown
IT Managers & Engineers
Falls Creek, Australia
www.mountainman.com.au/software

Received on Tue Nov 22 2005 - 20:07:47 CST