[starlogo-users] patch and global variables...
toontalk at googlemail.com
Mon Mar 12 18:48:41 EDT 2007
But one can argue that if one starts with celluar automata and then add
agents which are a fundamental generalization (since they can move unlike
cells) that you don't need the cells anymore. I'm not arguing against
celluar automata, only that they aren't needed in an agent language. Why
have two fundamental concepts in a language in which one is subsumed by the
Regarding giving up agents instead and sticking to pure cells:
"*Beware of the Turing tar-pit in which everything is possible but nothing
of interest is easy*" -- Alan Perlis.--
On 12/03/07, Kent Paul Dolan <xanthian at well.com> wrote:
> "Ken Kahn" <toontalk at googlemail.com> wrote:
> > Patches are too special purpose to be part of a language.
> That would be a rather large surprise to the
> software devotees of cellular automata, constructive
> solid geometry modeling, and certain flavors of
> genetic algorithms, all of which base themselves on
> a regular grid of unmovable elements-with-attributes
> which typically for the first and last disciplines
> communicate only with nearest neighbor cells or
> Such "computing with a grid of patches" is so
> important that several reductions of it to hardware
> implementations have been done over time.
> von Neumann proved in theory, and others since have
> reduced his proof to practice, that a cellular
> automaton "patch grid" can have a universal computer
> built atop it. Thus, there is in concept no
> programming problem that a patch grid will not
> suffice to solve. It is "agents" that are the
> additional complexity in the computational model,
> not patches.
> starlogo-users mailing list
> starlogo-users at media.mit.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the starlogo-users