Re: OT: A bit of Sudoku

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Fri, 13 Jun 2014 11:10:25 +0100, Mike H wrote:

> In message <ourmv.598292$hv6.317046@fx17.iad>, Shark8 
> <OneWingedShark@gmail.com> writes
>>
>>  -- Definition of "=" returns true when both sides are NULL,
>>  -- returns false when only one side is NULL, and behaves as
>>  -- normal string-equality when neither side is NULL.
> Is this an example where Dmitry's three state logical (True, Uncertain, 
> False)  might be useful?
> See <http://www.dmitry-kazakov.de/ada/intervals.htm>

As much as I wished, I can take credit for tri-state logic. It was known
since first third of XX century.

The extension is four-state Dunn-Belnap's logic (true, false, uncertain,
contradictory), of which natural application is for run-time check
conditions. [*]

Run-time checks could be inconsistent, e.g. when two sources contradict
each other. A trustful (AKA gullibility) combination of the sources (say
from two independent packages) yields "contradictory". A conservative (AKA
consensus) combination does "uncertain".

Interval relation use only conservative combination, thus tri-state logic
is sufficient for them.

----------------
* Reasoning based on lies, so to say.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de