Title: Message



You 
said it:  an evolvable system can accrete new features, although I am 
uncertain if novelty is required. 
 
However, to be evolution in the strictest sense, it must be a feature 
which subsequent copies or inheritors 
of the 
system propagate.  I'll call that 'viability' meaning this is a feature 
that will cause this system 
to be 
selected more often than siblings/versions without it,and therefore 
increases the frequency of 
occurrence of the system.  Evolution deals with populations.  
One should also point out as Jarron 
Lanier 
has, the time scale over which biological evolution works.  Anyway, let me 
pose the following 
qualties because they characterize control types:
 
1.  Random evolution:  a feature is added by a mistake.  
May be viable.
2.  Directed evolution:  a feature is added by 
intension.   Is viable or else is a mistake.
3.  Emergence:  a feature is added by a coupling of intensions 
and/or mistakes. Is viable by definition.
 
BTW: have you considered 
 
o  what happens if one routes values to the frequency and occurrence 
indicators of 
    a schema or a schema of 
schemas?
 
o  that your *related to* is a typed URI to a resource and you are 
delving into Page Rank as a 
    selector for services?
 
len

  
From: Roger L. Costello 
  [mailto:costello@m...]

  The ability of a system to 
  dynamically adapt and evolve to support capabilities that were never imagined 
  by the system designers ... ah, now that's what I want out of a 
  system!
   
  So, why do I ask [about the distinction 
  between extensible versus evolvable]?  Answer: I want us to 
  (colllectively) identify the fundamental differences and characteristics of an 
  extensible system versus an evolvable system.  
  
   
  Comments?