<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>If I haven't already beaten this horse into the ground:</p>
<p>The conceit of the program here is that rather than doing a
binary quad-tree decomposition of the 2D "Life" grid, we instead
decompose it into a denser quad-tree hierarchy based on an N-1xN-1
kernel. The point of this is to capture *all possible*
translations at *every scale*. The idea that drives this for me
is that by hashing these patterns (same as Gosper's original)
*should* yield the most compact state description of whole system
at any time step and therefore also across the evolution of the
system?</p>
<p>The bar chart on the right is green==number of hash hits and
red==number of hash misses for each scale 3x3...NxN.<br>
</p>
<p>There is theoretically an upper limit of the complexity of such
systems which is 2^(NxN) or for my 32x32 grid (anything much
larger bogs badly) 2^1024 (Yuuuge!).</p>
<p>This code (run in P5.js in my browser because I'm too lazy to
update the editor/runtime environment on my 15 year old computer?)
seems to exhibit the behaviour I would expect across many runs.(P5
is to me as NetLogo is to Guerin).</p>
<p> In the first pass, for example, the NxN has has precisely 1
hash entry while the exponentially increasing number of possible
entries is still sparse down to the smallest examples (3x3, 4x4,
5x5, etc). In *most* runs I've made the 3x3 saturates at 2^9=512
but sometimes it achieves quiescence or oscillation without fully
exploring all 512. The majority of larger scale patterns don't
even approach saturation (the numbers are just too large to
achieve in reasonable time but also the total computational space
(32x32x1bit) is not large enough to support very much exploration?</p>
<p>Next steps might include inverting the program to do recursive
computation rather than simply recording/measuring the patterns as
they unfold. A side effect of this would be to allow for
arbitrarily large patterns, not restricting the computation to a
pre-defined real-estate (e.g. 32x32)?</p>
<p>In spite of being a groupie to CA (and many forms of
computation/simulation) I have not really studied the larger range
of analyis of "Life" as Hashlife facilitates. I believe Guerin
turned me on to an example of the implementation of a glider-gun
based emulation of a conventional register/gate computer which
then implements hashlife (recursively), being "hashlife all the
way down"? It is possible that my N-1 kernel concept converges
with the N/2 quadtree of Gosper for "most" computations? <br>
</p>
<p>inquiring minds want to know.</p>
<p>(PS) been listening to Judith Rosen as I worked on this... I do
hope we can cajole/support Glen (or another) to engage... I am
also hammering on the recursive version but stalled because
multi-tasking with Rosennian and death/disposal and Fascist v
Oligarchical maunderings...<br>
</p>
<p><br>
</p>
<img src="cid:part1.Lm96J6ZV.vq40c79b@swcp.com"
moz-do-not-send="false" width="798" height="402">
<blockquote type="cite"
cite="mid:3ea36538-a208-4782-91dd-c2576ecc6b67@swcp.com">
<p>My weak grasp on the root of the reversibility argument is that
a great deal of the "information" that is generated in a
computation is in some sense extraneous? I wish I knew more
what I meant by that.</p>
<p>I took this opportunity to ride one of my hobby horses down the
road a little: a variation on Gosper's HashLife which is
designed to support a measure of computational complexity.
I'll elaborate under separate cover.<img
src="cid:part2.0ztF6RK4.dxcn6OJx@swcp.com"
moz-do-not-send="false" class="" width="853" height="431"><br>
</p>
<blockquote> </blockquote>
<p><br>
</p>
<blockquote> </blockquote>
<br>
<div class="moz-cite-prefix">On 1/18/25 5:03 PM, steve smith
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:db727e93-0c98-436b-bdfb-42dec31dfd3d@swcp.com">
<meta http-equiv="Content-Type"
content="text/html; charset=UTF-8">
<p>Pieter -</p>
<p>Good find. It lead me to Vaire and then to the Sandia/ABQ
work of Michael Frank who left to join/found Vaire this
summer? It is possible that my renewed interest in
reversible computing might have been triggered subliminally by
some reference to both/either? </p>
<p><a
href="https://techcrunch.com/2024/07/01/vaire-computing-raises-4-5m-for-reversible-computing-moonshot-which-could-drastically-reduce-energy-needs/"
moz-do-not-send="true">https://techcrunch.com/2024/07/01/vaire-computing-raises-4-5m-for-reversible-computing-moonshot-which-could-drastically-reduce-energy-needs/
<br>
</a></p>
<p><a href="https://vaire.co/" class="moz-txt-link-freetext"
moz-do-not-send="true">https://vaire.co/</a><br>
</p>
<p>I thought I'd been triggered by the combination of the
demands of AI and on data centers (my daughter closed her gym
of 10 years to take a job in a data center development startup
a year ago... ). <br>
</p>
<p>My inability to attribute such things, parallels that of LLMs
(or more generally transformer models)?</p>
- Steve<br>
<br>
<fieldset class="moz-mime-attachment-header"></fieldset>
<pre class="moz-quote-pre" wrap="">.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-..
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom <a
class="moz-txt-link-freetext" href="https://bit.ly/virtualfriam"
moz-do-not-send="true">https://bit.ly/virtualfriam</a>
to (un)subscribe <a class="moz-txt-link-freetext"
href="http://redfish.com/mailman/listinfo/friam_redfish.com"
moz-do-not-send="true">http://redfish.com/mailman/listinfo/friam_redfish.com</a>
FRIAM-COMIC <a class="moz-txt-link-freetext"
href="http://friam-comic.blogspot.com/" moz-do-not-send="true">http://friam-comic.blogspot.com/</a>
archives: 5/2017 thru present <a class="moz-txt-link-freetext"
href="https://redfish.com/pipermail/friam_redfish.com/"
moz-do-not-send="true">https://redfish.com/pipermail/friam_redfish.com/</a>
1/2003 thru 6/2021 <a class="moz-txt-link-freetext"
href="http://friam.383.s1.nabble.com/" moz-do-not-send="true">http://friam.383.s1.nabble.com/</a>
</pre>
</blockquote>
<br>
<fieldset class="moz-mime-attachment-header"></fieldset>
<pre class="moz-quote-pre" wrap="">.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-..
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom <a class="moz-txt-link-freetext" href="https://bit.ly/virtualfriam">https://bit.ly/virtualfriam</a>
to (un)subscribe <a class="moz-txt-link-freetext" href="http://redfish.com/mailman/listinfo/friam_redfish.com">http://redfish.com/mailman/listinfo/friam_redfish.com</a>
FRIAM-COMIC <a class="moz-txt-link-freetext" href="http://friam-comic.blogspot.com/">http://friam-comic.blogspot.com/</a>
archives: 5/2017 thru present <a class="moz-txt-link-freetext" href="https://redfish.com/pipermail/friam_redfish.com/">https://redfish.com/pipermail/friam_redfish.com/</a>
1/2003 thru 6/2021 <a class="moz-txt-link-freetext" href="http://friam.383.s1.nabble.com/">http://friam.383.s1.nabble.com/</a>
</pre>
</blockquote>
</body>
</html>