[Edu-sig] RE: [Tutor] Off topic musings

[David Scherer]

> Also, since a computer is a tool that we have created, we 
> also control the "laws" which govern the design, operation 
> and use of this tool. As I see it, laws are generally 
> universal constants - a way to explain and predict the 
> behavior of natural forces. Laws mean nothing when you can 
> change them at will. We define the significance of bits and 
> bytes, we create the programming languages which we use to 
> create software, etc. We can also change those things, 
> meaning they aren't held constant. The only 'laws' I see as 
> applicable to computer science are those that it uses in 
> creating these tools - i.e. electricity and conductivity.

There most certainly are universal laws that apply to computers,
regardless of how they are created(*).  For example, there are problems
that cannot be solved in general by any computer program.  Any computer,
given unlimited memory, can simulate any other computer.  There is even
an absolute definition (Kolmogorov complexity) of the information
content of any object, though it is not computable.

If you accuse Computer Science of being insufficiently "unified", I
can't argue with you, because I don't really know what that means.  But
there are most certainly universal laws of computing!

Look here for one of the weirder consequences of this type of work, and
probably the best paper title of all time:

http://citeseer.nj.nec.com/hutter01fastest.html

Dave

(*) Obviously it cannot be shown mathematically that every possible
computer obeys these laws, but there are many examples and no
counterexamples, which is all that we can say for any scientific theory.
Everything that we call a "computer" today most certainly does obey
them.  Anything based on logic gates and memory will obey them.
Probably anything based on known laws of physics obeys them, since as
far as I know the known laws of physics can be simulated (very slowly!)
on a Turing machine.