Re: How to cope with missing values - NULLS?

From: Peter Koch Larsen <pkl_at_mailme.dk>
Date: 22 Apr 2003 15:40:49 -0700
Message-ID: <61c84197.0304221440.32113f01_at_posting.google.com>


"Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message news:<b83u65$1imu$1_at_gazette.almaden.ibm.com>...
> "Peter Koch Larsen" <pkl_at_mailme.dk> wrote in message
> news:61c84197.0304220818.3050fbdf_at_posting.google.com...
> [ snip ]
> > There is not much of a difference the way I look at it and it really
> > bugs down to how you feel the "=" operator should work.
>
> Indeed it does really boil down to how equality works..
>
> Quite simply I take it as a axiom that
> x = x
> for all possiable values x
>
> if you want to break this by introducing something that is not a value, then
> be my guest, but don't expect many others to follow you.

Should I take this sentence as meaning that if only SQL defined NULL = NULL as true everywhere (i nearly wrote "in a consistent manner", but this must be superfluous) then the NULL-type would not be so bad - at least in special places such as the outer join (I cant think of any other place where nulls do occur naturally)? If not, I would be very interested in hearing how your union type differs from the above, and if it does not what missing properties you would like these union-types to possess.
>
>
> > One of my thoughts have been that it might be a good idea to introduce
> > an equivalence operator (we could call it "~") behaving the same way
> > as your operator does (UNK ~UNK = TRUE); I've never heard this idea
> > proposed anywhere else.
>
>
> Indeed you could, but please use it as a founding operator of your own system
> of mathematics and stop bothering us with such stuff.
>
> If you are interested in what (some of) the rest of the world thinks about
> equality, try following the link Costin gave out recently
> http://www.nuprl.org/html/NaiveTypeTheoryPreface.html

Interesting. I will check it out.
>
> Regards
> Paul Vernon
> Business Intelligence, IBM Global Services

/Peter Received on Wed Apr 23 2003 - 00:40:49 CEST

Original text of this message