|
|
Jan Burse wrote:
> Hi
>
>
> 3-valued logic has very nice applications in
> logic programm, in database, etc..
>
> For example SQL is based on 3-valued logic,
> you have there TRUE, FALSE, UNKNOW, the later
> arising if null values are used.
Yes, I'm more or less familiar with this application. Several years
ago when I was working with 3VL, I was directed to a series of articles
which advanced various arguments in favor of and opposed to the use of
3VL in SQL. The proponents argued that it worked. The opponents argued
that the theory was weak and would lead to unsound implementations and
erroneous answers to queries. After I had made sufficent progress in
my own studies, I returned to these articles and found, in short, that
both proponents and opponents were partly correct. The 3VL used in SQL
is a fragment of what I have, and in many versions a poorly understood
and poorly implemented one at that.
I' m entirely unfamiliar with logic programming, so your second
example means nothing to me. About all I can say is that at least one
Kleene logic described in the references in MVL can also be taken as a
fragment of the enhanced Lukasiewicz logic I have been working with.
But these applications don't invalidate the point I was trying to
make. These (and most other MVL's) lack conditionals that are
sufficiently well behaved to permit ordinary deduction. It would be a
far stetch to call them *modal* logics.
|
|