>> I had considered this approach [using illegal values to indicate
>> missing values] but i keep hitting the mental block
>> which tells me that a missing value is not a value and that using a
>> default value "feels" wrong... [ Of course I will use defaults where
>> they are appropriate however :) ]
>I've found 3- and 4-valued logic quite sensible, but that may be an
>effect of the many years I've worked with Prolog (a logic-based
>programming language) as a database manipulation language.
>The problem is that SQL doesn't really support missing values.
>Paraphrased from an article which appeared about 10 years ago in an ACM
>periodical: "If programmers became able to program in logic, we would
>discover that programmers are not logical." That year I learned from
>direct observation that the writer was correct and optimistic.
I would suggest not using any values that would correspond to a clue that
the database is corrupt. For example, find out what the database thinks
is the value of a space that is misdefined as a number, and don't use that.
This hint was just about useless when database engines controlled all
metadata, but with the advent of metadata distribution and different
ways of accessing archived data these errors are coming back again.
Columns of 8224 was always a clue that something was wrong in the DMS-500
>1801 California (on assignment from Analysts International)
>Denver, CO 80202
These are my opinions, not necessarily those of Ross Systems, Inc. <> <>
%DCL-W-SOFTONEDGEDONTPUSH, Software On Edge - Don't Push. \ V /
panic: ifree: freeing free inodes... O