[Reader-list] Alternative software???

[Supreet Sethi]

On Sun, Dec 01, 2002 at 04:33:57PM -0500, Are Flagan wrote:
> On 11/29/02 18:53, "Pankaj Kaushal" <pankaj at sarai.net> wrote:
> 
> > ***Sorry fellow earthlings I give in to the troll***
> > 
> > Dear Are,
> > Sorry for the interruption but as a student of Theory of
> > computation, I am sorry to say neither I or Turing will
> > agree to the first paragraph of your mail
> >
> 
> Of course he would not agree. See the notes again, especially his sixth
> objection on Lovelace. Then again, I do not agree with Turing (nor do I
> think he agreed with his earlier self in the end).

First of all what turing thought of himself, his brother his sisterhis 
wife or his work is not the subject. It does not contribute to qualitative 
analysis of computer science wheather he was gay or hetrosexual.




>  
> > 
> > First of all the question is not about either hardware or software
> > or that software is dependent on hardware. the later is just an
> > assumption people make or are taught by other people who dont understand
> > computing very well. The mail motive is to solve a problem and
> > if a turing machine can solve that problem then it is logicaly solvable
> > and then the implementation of it can be in hardware or hardware dependent
> > software or hardware independent but software dependent software.
> >
> 
> What you are missing is the Turing breakthrough in modern computing -- the
> interplay between the problem, the algorithm, the Turing machine,
> software/hardware and the Universal. This can of course be configured in any
> of the ways you mention, but this discussion, along with the essay, was
> aimed at an elucidation of software, that is to say primarily one
> configuration of the set up. Turing never really made a distinction between
> hardware and software, except by adding the prefix Universal -- Turing
> Machines within a Universal Turing Machine -- suggesting that they are of a
> similar order, but stacked like Russian dolls. This point is reiterated in
> the essay by making some reference to the hazy distinction between software
> and hardware, while focusing on their limits and the operative desires and
> effects, so to speak, of software. Or, if you like, the function of your
> tautology -- "if a Turing Machine can solve a problem then it is logically
> solvable." To which my essay incredulously attempts to ask: because the
> Turing Machine (software/hardware) told you so?



The universality of turing machine does not mean "a" turing mahcine could 
solve all the problems. Turing mahcine is a simplistic model on 
super-scalar structure called computing is based. Just looking at one 
neuron in human brain you can't figure out how many things human brain can 
do, similarly you cannot just assume that this simplistic model, yet 
proven model cannot deal with complicated scenearios. 

A neuron also performs almost the same function as a regular transistor 
would. It switches on/off.




>  
> > There are certain problems which considered undecidable thus, unsolvable by
> > computers for example the problem that the set of all functions f:N->N is
> > unsolvable by turing machine.
> >  
> > Again it is foolish to assume that software is dependent on hardware or that
> > the universal principles of computation are laid down by hardware, hardware
> > design or limitaitons of hardware components.
> >   
> > There is a concept called nondeterministic turing machine just as
> > finite automata is allowed to act nondeterministcally. a nondeterministic
> > turing machin can produce two different outputs to the same input.
> >  
> > do you have any clue to what you are writing interpreting
> > imssinterprating.
> >
> 


what you missed is that any non-deterministic automata can be reduced into 
a deterministic automata.






> No clue (so let's ignore that angle). There are many more problems
> unsolvable by a computer, such as chewing gum and walking at the same time
> (but it is nicely simulated in multitasking). 

It could be just that simulated multitasking, how do you know that human 
brain is not using the same algorithm of time-shared multitaksing in its 
various sub-"computers".





The prospects of understanding
> computing (or software or hardware) cannot belong to math and logic alone,
> unless these are also recast within their own social and cultural and
> philosophical histories. If computing is only understood as a successful
> program -- invoked here in the widest sense but with reference to its
> application -- that relies upon certain solutions (arrived at through math
> and logic), understanding then resides solely with the machine that _runs_
> them, brings them to life.

Math and logic if you notcie tend to be very dependable. A cartesian 
product of two sets would be same. Same euclidian geometry forms the basis 
of any new discovery. Socail /cultural/philosophical histories are 
subjective. Every next guy would interepret it diffrenetly. Besides I 
think human brain underneath all its complexities relies on same math and 
logic. In macro terms may be nobut at micro level for synaptic shifts same 
AND OR NOT XOR logics form the basis of information storage and retireveal 
system.




 Post Heisenberg, the Russian attempts at building
> trinary computers, and the speedy benchmarks of photons rather than
> electrons, one could indeed say that it is foolish to suggest that there is
> any consistency or limitations to hardware or software, or any set
> dependency to their relations. However, anyone who has ever pointed at a
> stack to set either a true or false value/voltage would acknowledge that
> there is a dependency, and I have yet to run an algorithm that does not
> resolve its truth table, speaking here of limitations (my computer has
> crashed on occasion, though). What I vehemently seek to contest is the
> built-in arrogance that seeks refuge in the operations of the machine, in
> its concepts and functions, to render any other interpretations than those
> proffered by its devoted

Unversality of turing machine BTW was never asserted by turing himself. 
Its something called church's thesis which asserts that turing machine can 
solve any algorithm solvable by a real computing machines. Which I think 
is a right assertion. 

What i understand of a turing machine is the fact that any computation job 
can be reduced to a model where a left to right tape is moving and a 
reading head can interprete and take action on the basis of it. If you 
look at this model closely it looks where similar to a neuron in a human 
brain.




if you see carefully there is arrogance every where. you have to prove 
what you are saying by building an alternative model. If you coul build a model 
which supercedes turing machine as a mathematical model for a computer, I 
would probably be the first one to write a compiler based on it.

BTW is'nt there a arrogance in you when you say Free Software should be 
doing this , not that or the other.

Evolution follows a very organic pattern, see turing machine as a species 
of computers, it is a proven model of doing things. Its on the top of the 
food chain. That does not mean there is no other model. There is already a 
quantum compuetr build by IBM. If it proves to be a better way of doing 
things it would becomes part of everyday household . Who know a million 
year from now a computer would be telling another computer that his human 
slave is lazy and consumes to much energy. That would be sweet revenge I 
think.


> 
> -af
> 
> _________________________________________
> reader-list: an open discussion list on media and the city.
> Critiques & Collaborations
> To subscribe: send an email to reader-list-request at sarai.net with subscribe in the subject header.
> List archive: <https://mail.sarai.net/pipermail/reader-list/>

-- 
supreet at sdf.lonestar.org
SDF Public Access UNIX System - http://sdf.lonestar.org