David,

D> I think that I agree with an earlier post that the use of the word
D> "NULL" was a poor choice for the standard. However, it is a well
D> defined standard with expected behavior across all RDBMS systems. I
D> don't know about relational math, but the real-life practice is
D> very practical and not nearly as awkward as your posts seem to
D> suggest.

I have posted just once and in addition it wasn't meant to say that I
am against NULL. I am in peace with it since quite some time. :-)

What I meant to say was that a way to devise a "special" value would
have done a better job of expressing unknwon quantities. This is an
opinion I happen to agree with but I have in no way invented it.
Actually I have borrowed it from C. J. Date himself.

www.dataphor.com has a good implementation (even though it's in the
middle of an endless beta cycle) and some papers describing the
foundations that once convinced me and might convince you too.

D> There is room for latitude in implementation, but the behaviors of
D> elements (like NULL values) are well defined by a cross platform
D> standard.

Couldn't agree more. I use and take advantage of NULL myself. It's
just that everything can be improved. ;-)

D> Rather than taking this up with the Firebird support list, take it
D> up with ANSI and the ISO, who defined the standard that Firebird
D> implements.

I'm not taking it up with anyone, and BTW Firebird's faithful
adherence to the SQL standard is one of the reasons I use it.
I have just added a 5 line OT comment to a thread that was drifting OT
anyway. I can never resist. :-)

Ciao
--
Nando Dessena
mailto:nandod@...