Lists Home |
Date Index |
-----BEGIN PGP SIGNED MESSAGE-----
> -----Original Message-----
> From: Bullard, Claude L (Len) [mailto:email@example.com]
> Sent: 20 May 2002 14:26
> To: 'Bill de hOra'; firstname.lastname@example.org; 'Eric Bohlman'
> Cc: 'xml-dev'
> Subject: Intelligent Adaptive Systems (WAS RE: [xml-dev]
> Objections to / uses of PSVI?)
> Adaptation requires extra resources or very precise coordination
> with the environment as it changes. The issue is the ability
> to identify a pattern and predict changes. These are markov
> problems. If thriving in an environment requires instrumenting
> the environment to detect emergence, it is quite straightforward
> for a well-adapted system to evolve with the environment.
Adaptation first requires the ability to adapt; trite as it may
sound, that implies inbuilt redundancy.
> Are you discussing intelligent adaptation?
Not as much as engineering in degrees of freedom to allow
adaptation; the behaviour might look intelligent from a distance,
but it's largely a consequence of not being brittle (brittle things
are in a sense, stupid). If all the subsystems are maximised
instead of optimal there's little or no room left to change and no
execess capacity to invest. It's why we build buffers between
computer processes; we typically tune them for efficiency not
De Marco's book is worth reading, no surprise there.
Bill de hÓra
-----BEGIN PGP SIGNATURE-----
Version: PGP 7.0.4
-----END PGP SIGNATURE-----