<HTML> <HEAD> <TITLE>Archives of ctltheory for Sep15-98</TITLE> </HEAD> <BODY BGCOLOR="#FFFF00" LINK="#FF0000" VLINK="#FF0000"> <!-- <CENTER><IMG SRC="http://your.server.com/your.banner.gif"BORDER=0></CENTER> --> <P> <CENTER> <TABLE BORDER="0"> <TR> <TD ALIGN="CENTER" COLSPAN="2"> <FONT SIZE=+2><B>Archives of ctltheory for Sep15-98</B></FONT><BR><HR></TD> </TR><TR> <TD ALIGN="LEFT"><B>Tue, 15 Sep 1998</B></TD> <TD ALIGN="RIGHT"><B>Vol. 1.13</B></TD> </TR> </TABLE> </CENTER> <P> <A NAME="Contents"><STRONG>Contents</STRONG></A> <BR> <HR WIDTH=50% ALIGN="left"> <BR> <UL>01. <A HREF="#SUBJECT01">PPM</A> (Trevor)<BR> <A NAME="SUBJECT01"></A><A HREF="#Contents">Return to Contents</A><P>----------------------------------------------------------------------<P>Date: 15 Sep 1998 18:23:55 -0400<BR> From: Trevor <aki@xs4all.nl><BR> Subject: PPM<P><P> PPM INFORMATION RELEASE:<P><BR> Recently, a "Perpetual Pattern Machine" was terminated (due to<BR> reasignment of computational facilities) after around thirteen hours of<BR> apparently successful continual operation.<P> Auto-morphogenetic and other anti bit-rot processing techniques were<BR> used in a series of experiments involving small ecologies of<BR> Movable Finite Automata to study resistance to entropic decay in<BR> conditions of intense feedback.<P> The systems investigate the mechanism of continual (non-biological,<BR> i.e. non-sexual) evolution in closed systems. Once started these<BR> processes are completely deterministic.<P> Contining the trend towards simplicity, the current series is<BR> reduced to parallel sets of basic processesing units variably scanning<BR> through the image in an ordered manner (Space-Time Processing).<P><BR> A major problem is the question of proof: Due to a random start<BR> (big bang), each "existance" is one of an extremely large set of<BR> possible trajectories and each trajectory which has not (yet)<BR> degenerated into chaos or repetition may of course do so at any moment<BR> in the future. Chaos (as a subjective phenomenon) is fairly easy to<BR> recognise -but how does one watch a large number of "existances" for<BR> an infinite length of time to see if they are repeating (how does one<BR> compare an image with one from three days earlier that looked a bit<BR> similar)? How similar must images be in order to be concidered<BR> "repetitive" (or how different to be concidered as evolving)?<P> Initially as a practical measure (in order to improve performance on<BR> a fairly slow Amiga 3000) most internal integer functions were<BR> represented<BR> by look-up tables. This gives the additional advantage that the<BR> underlying<BR> conceptual model of the "Turing Machine" becomes more explicit and<BR> therefore more easilly integrated into a "Parametrical Space" model<BR> so the process becomes more obviously concerned with the transformation<BR> of closed ordered sets.<P> The look-up tables also make more explicit the nature (dimensions) of<BR> the spaces involved as well as the transformations concerned and so may<BR> therefore form a tentative step towards a "fungible" computer.<P><P> Institute for Autogenetics<BR> and Metamorphology<BR> Amsterdam, September 1998<P><P><BR> </UL> <HR WIDTH=30% ALIGN=left> <A HREF="mailto:ctltheory@lists.timesup.org"> Post a message to the list</A><P> <A HREF="index.html">Return to Sep. Index</A><BR> <HR WIDTH=30% ALIGN=left> </BODY></HTML>