Re: Forward vs. Reverse Engineering

From: HARNAD Stevan (harnad@cogsci.soton.ac.uk)
Date: Tue Jun 04 1996 - 10:35:31 BST


> From: "Wright, Vicky" <vmw195@soton.ac.uk>
> Date: Fri, 24 May 1996 10:01:53 +0100 (BST)
>
> Forward engineering is when devices are designed and built to perform
> certain tasks using basic already acquired knowledge, for example
> planes that fly. Reverse engineering is the opposite of this process,
> the devices for example, people, are already built and the "engineers"
> have to find out how they are designed, this is the case in cognitive
> psychology - the attempt to try and "explain the mind" as cognitive
> psychologists try and give a causal explanation of our mental and
> behavioural processes.

The stuructures and processes that give rise to our CAPACITIES.

> Cognitive psychology forms theories which are
> then tested and refined by observing behaviour and brain processes in
> order to generate models i.e. artificial devices which are also capable
> of performing the same tasks so we know it can be done.

And gradually scaling up to the Turing Test, where it can do EVERYTHING
we can do. As the capacity gets broader and matches our own more
closely, the hope is that the number of alernative ways in which the
exact same thing could be accomplished will shrink, leaving only the
"right," way, which is our way, and the same one the Blind Watchmaker
found.

> Reverse
> engineering is an example of bottom-up processing. All the "why"
> questions in reverse engineering have answers. Reverse engineers have a
> great deal of self knowledge and although engineers sometimes make
> mistakes in their designs and overlook retrospective shortcuts, the
> "default assumption must be optimality as if reverse engineering can't
> assume that there is a good rationale for the features they observe,
> they can't even begin their analysis." - Dennett.

Kid-sib would not understand that quote: Explanation is needed. For an
A, integrate with bigger themes too, such as computation vs. analog
processing and neural nets, or the issue of consciousness, or
causality.



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:44 GMT