the logic of the infinite

Kyle Lahnakoski
Thu May 23 17:50:02 2002

jeremy wrote:
> 1. From my initial and partial understanding of Arrow
> information theory, Lisp, and such, the basic idea is to
> abstract the underlying structure and function of real-world
> entities and processes as much as possible. Maybe this isn't
> a good way to put it. Anyone care to correct me or clarify
> this point?
> In order for this to be effective, our basic abstraction, our 
> base class, our informational atom (kernel?), needs to be
> extremely flexible and able to model the form and behaviour
> of any entity whatsoever. 

When you add flexibility/expressivness to one aspect, you reduce
flexibility/expressivness of another.  The desire to model a complex
thing will not be helped (much) by having a flexible primitive.  

In any case, from an informational point of view, the concept of the
reference is the simplest primitive for building models of nature.

Of course data modeling is not everything, and the whole issue of
*action* still has to be dealt with.  These issues are  handled by
functional languages and the lisps.

Even when you have a flexible modeling environment, you still do not
have anything useful.  You can see that both Lisp and Smalltalk are good
foundations, yet have had so much more work put into them than just
designing their semantics.  I am referring to the modules (GUIs, text
editors, web protocols, development environments, etc.) that make a
system useful.  Even still, Lisp and Smalltalk are not close to
implementing the Tunes aspects (like verification and migration).  These
require real work that is not helped by fiddling with representation

I suggest working on these problems directly without concern for the
environment you build it in.  The hard work of clarifying the necessary
aspects of a solution can be done now, and easily migrated to a Tunes

> We need a framework that can express the nature of reality,
> which is of one infinite being assuming an infinite diversity 
> of forms and behaviours. The basic abstraction would then
> need to be a model of infinity.  ?what do you think?

We know that the "infinity" of nature can not be captured inside any
static specification.  The best we can do is have a system that will
accept new contextual axioms to help us model new things.  When you
consider an extensible system like this, I think you will agree that it
is just as useful, and definitely easier to build, than some "infinite"

> 2. My approach to programming is obviously top-down (or
> rather inside-out), big-picture and quite abstracted from the
> mechanics of computer hardware. This is a huge shortcoming
> for me and I would love to hear what people with a more
> realistic understanding of computer systems have to say about 
> my silly notions. Your comments will help me refine my views
> and bring them more in line with the underlying philosophy of
> the Tunes project.

I think you are thinking/imagining at too high a level.  No useful
conclusions can be made when you are not bounded by the mess we call
reality of computing.

Most of your quotes do not apply because they talk of the infinite (read
unexplainable) and its relation to the finite (read explainable). 
Wonderful concepts for philosophy, but inapplicable for developing, or
understanding a real system.  To build Tunes, or something like it, one
must stay in the realm of the finite (read explainable); contemplations
of the infinite are unnecessary.