Well, there are two sides:

1. The body analogy is only showing that a very very complex system like a
human being,
build out of small objects - called cells - CAN function on basis of
messages and specialised behavior and fight for 80 years the power of
osmose, Period.  

2. Introduce intention as explanation for the direction of this
specialisation? 
(Please not again: What intention did that amoebe have? What is the
intention of that embryo to have branchies? What kind of rational behavior
could explain the motives of.... etc.)

Back to topic: So if you CAN build very complex systems with objects, How do
you keep overview as normal human being during this complex building proces?

   
We all know that we lose very quickly the overview (magical number seven
plus or minus two, do you remember?)  
We also know that experts don't remember more, but organise details of the
problem in schemes and patterns (A.D> de Groot: thinking of the
chessplayer.)
We also know that when you lose the overview mathematical prove of
correctness of complex designs is not possible: Knuth:"Watch out, I did
prove it is correct, but I did not test it."
And if the reading of Gödel Esher and Bach did not convince you, then there
is always that sucking feeling of chaos theorie.)

What then:

Top down design? No
Bottom up? NO
Team approach?
Waterfall project-approach to fight complexity? NO (who tells it our
managers?)
RAD? Going extreme?

In his exposition about fast&good reading Walsh did give the answer: do all
together!

(It is the way you and I solve a problem: In the beginning you scope the
problem, then you come up with possible solutions, then you work on one - or
better on some (deBono advice) - and evaluate these and make plans for the
next round. BUT MOST IMPORTANT OF ALL: Right form the start you already
think of the detail-parts of the problem AND UNTIL THE END You keep scoping,
refining the overview part of the problem.  George Polya 1957)

How can you organise all this. I do NOT say automate the solution, just
organise this design process of skipping between the overview with the eye
of an eagle and working deep in the durty details ?
I think that OO and Smalltalk did try to do that as first: organise
compexity in a browser.
Do not blame them that the tool becomes so complex: it has to cope with
compexity.

(UML and design approaches like RUP, and Rosenberg with Iconix & the Agile
people are trying to beat them. I still think that using CRC-cards for the
first cut, before going in Squeak would be a good start: When does CRC
show-up again as part of Squeak?)


 


> -----Original Message-----
> From: Justin Walsh [mailto:jwalsh at bigpond.net.au]
> Sent: donderdag 11 oktober 2001 23:12
> To: squeak-dev at lists.squeakfoundation.org
> Subject: Re: Message oriented genetics
> 
> 
> Good reply! But, would it not be better to discover the 
> (rational) behavior behind intention first? Might it not save us a lot
wasted detail?
> ----- Original Message -----
> From: "Andres Valloud" <sqrmax at prodigy.net>
> To: <squeak-dev at lists.squeakfoundation.org>
> Sent: Friday, October 12, 2001 1:43 AM
> Subject: Re: Message oriented genetics
> 
> 
> > Hi.
> >
> > > > Sometimes I wonder what would happen if somebody came up with
> > > > Message Oriented Quantum Mechanics, Message Oriented Physics or
> > > > Message Oriented Genetics...
> > > If I try to explain to someone the idea of objects and messages,
> > > they find it very difficult and abstract:
> > > Then I do the body analogy <snip>
> > > We do not know what really triggers this specialism: biochemical
> > > messages? (or.. gravity?... relative location?)
> >
> > It would be best to spell out the behavior which we can attribute to
> > intention.  Implementation details will get better over time as more
> > intention is discovered.
> >
> > Andres.
> >
> >
> 
> 
>