Approach for implementing TUNES

Wed, 19 May 1999 15:58:12 -0700

> I think you are getting a little ahead of yourself, or me at least! Your
line of
> thinking is similar to mine about my second generation OS. The problem
you face
> is that you are trying to do missuse deduction. The only thing deduction
can do
> for you is tell you about your premices. So when you run your program you

Well... All information it deduces will be based on premises, surely ...
But deductions at the end of a long chain can often look very unlike the

> learn about many abstract relationships inside your computer but few of
> will have any practical value. You need a method for generating a large

Don't you think there is great practical value to knowing a boot image for
TUNES (for a given type of machine), don't you think?

> complex system that can assimilate new functions and provide an object
> interface to the user. In other words what Tunes has been evolving into

System that can assimilate new functions ... object-based interface. Well
if this is the kind of system you want, then you should be able to a find a
boot image for it with my logic system (if it ever gets created).

> reciently is a quite promising attempt at creating a Strong AI. :) I have
> highest hopes that we will succede if we continue in the direction we are
> headed. 
> As you see this is why I have decided to make this kind of OS a second
> generation goal. In the mean time It is best to apply sound design
techniques to
> building a highly modular conventional OS similar to LiNuX but MUCH MUCH
> and more flexable for new types of applications and devices.
(cybernetics, 3D
> displays, Semi-autonomous robots.)

Well ... certainly a better OS could be good even if it is not as good as
what could be generated by my logic system (if it gets created). It
certainly would be a good idea to work on Brix or Retro in the meantime
since, as far as we know, this logic system might not work out...

- Brent Kerby ("Iepos")