Re: [OFFICIAL] Some FT background stuff (guidelines for writers) - LONG POST!

More ramblings!  (8-)

Allan Goodall wrote:
> 
> On Thu, 12 Feb 1998 02:15:28 -0500, Jerry Han <jhan@canoe.ca> wrote:
> 
> >An unmanned figher is only as good as its programming.  In a combat
> >scenario, where mission execution may depend on the unexpected, human
> >participation in the decision loop is essential.
> 
> There is human participation: the programmer. The programmer can be
> wiley and unpredictable; he just has to think of it ahead of time...

The programmer has to be good.	But, this raises the spectre of good
software engineering, which, so far, has proven to be an impossibility.
I'm a programmer, and I will say that programming is still more art than
science.  We can't take a measure of a program the way we take a measure
of a bridge or a starship and say it's going to work or not.  In fact,
there are some computability theorems (Halting Problem) which basically
state we can't predict the output of a program without running the 
thing.	If it's a large intensive software project, running it several
thousand times might be better, if you want to catch all the major bugs.

As for my decision loop argument; the programmer is too far removed to
be of use in the information-decision cycle.  There should be somebody
in the cycle who can intercede at a moments notice in the event the AIs
encounter something that wasn't programmed for, or in the event of 
program failure.  (It would be very bad if you lost your entire fighter
wing because your entire force used Windows Combat AI as operating
software.  (8-) )

J.

-- 
Jerry Han -  CANOE Canada - jhan@canoe.ca - http://www.idigital.net/jhan
  ***** Visit the Canadian Online Explorer! => http://www.canoe.ca *****
The opinions expressed are mine, and not necessarily those of CANOE
Canada.
	 "Is there no escape from the words that plague me so?"