New class of T machines...

[Billy Tanksley]

From: Alan Grimes [mailto:alangrimes@starpower.net]

>As much as I try to get beyond turing machines (preferring the lambda
>calc), I just had an idea regarding them: 

The two aren't well related.  However, Abstract State Machines conceptually
provide something of a link between them.

>Not all turing machines are the same.
>You can build a turing machine to do any function, In many cases the
>resulting machine can do ONLY that function. But still there 
>is a subset
>of those machines that are called *UNIVERSAL* turing machines. 
>These are
>such that they behave as interpriters for any language of your choice
>using any grammar. This language in turn can be used to evaluate any
>function that can be expressed in them, including interpriters 
>for other languages. 

This is true.

>My idea is that there exists yet another subset of these universal
>turing machines that could be called "universal universal turing
>machines" that operate as follows: 

This is trivially true: that subset is an improper subset.  All universal
turing machines are universal universal as well.

This is true by definition of the word 'universal'; if Turing and his
cohorts had not meant the word 'universal' they wouldn't have used it.

-Billy