ConwayLife.com

@alexv, https://git123hub.github.io/ReversibleLife
plan: CART can import Rule table
For me it was not most actual, because of mentioned converter and also due to problem with table format itself - it seems not adapted well for reversible cellular automata (see tables for WlAnt). On the other hand ...

PHPBB12345 wrote: is "CART" can import Golly table format it?

No, only *.tree format, but Golly older 2.4 distribution includes RuleTableToTree.exe converter table -> tree.

https://git123hub.github.io/index/
Thank you for reference to CART there. I just updated WlAnt directory to include "wireless snakes" mentioned somewhere in the thread. There is a new opportunity with non-totalistic rules in Golly 2.8, so it may be interesting to write here with new notation first ...

PHPBB12345 wrote:https://github.com/git123hub-group/wireless-ant

Slightly updated version (e.g., newer Golly *.rule format) is here
https://github.com/qubeat/CART/tree/master/delphi/WlAnt

https://git123hub.github.io/WlAnt.html
...

How does it work? I failed to enter and run something. I myself testing github page for reversible cellular automata related with WlAnt, but without any JavaScript demos yet
https://qubeat.github.io/CART/

PS. After all I find, how it works - I suppose ...

Just have known about a new game based on the WlAnt
http://ludumdare.com/compo/minild-57/?a ... &uid=40339

Alexander

pi_guy314 Any hope for a another bounded rule version?

Yes, simpily just remove the transition under square loop destruction on any existing bounded rule.

In fact, it is enought to change "n_states:14" to "n_states:15" (the next line after @TABLE
and rename file and rule, changing first line ...

For me the motivation was quite natural, because after looking on shapeloop I myself was interested,
if it is possible to get rid of some auxiliary states used as support for carriers of information about shape,
but now I did not quite understand about new version of Novoloop rule: if loops with ...

Here is minimalistic bounded version

@RULE 2armshapeloop-a
1 wire
2 left signal
3 right signal
4 forward signal
5 trigger right
6 idle right
7 converter
8 lead
9 edge 1
10 edge 2
11 starter
12 tempdel
13 delall
14 boundary
@TABLE
n_states:15
neighborhood:Moore
symmetries:rotate4
Var a={0,1,2,3,4,5 ...

Bigger loop won here
http://www.conwaylife.com/forums/viewtopic.php?f=11&t=1446#p13470
but not against smaller known 6x6 loop we are talking about. In my project I added
the loop to some competition and mentioned mutants suddenly appear, but they can
not win the 6x6 loop, yet may compete longer ...

I did not write that the project succeed.

Seems I missed something with the rule: is it possible in principle to have square or rectangular loop with
possibility to reproduce itself? All my tries to do such simple thing has failed.

Not the same behavior, but I should uncover a top secret. Both mutants from previous post appear in project to
win the terrible 6x6 pattern you mention.

This mutant has even more boring behavior

x = 10, y = 13, rule = shapeloop
3$2.ADCB$2.A2HC$2.AHIDCD$2.AHI2HD$2.AH2.HD$2.A4HF$2.4AEA!


PS. Yet, an interesting property of the loop is behavior after reflection. The resulting loop may not produce
own copies - it's producing other loop.

x = 14, y ...

It's happen I forgot turn off Golly and suddenly found pattern with 400K+ generations and two mutants
on a background window. I started competition between them and "SquareL" and they seem more
successive after 100K steps ... 200K steps ... SquareL still keeps two arenas. The real competition .

x ...

BTW, I have checked minimal modification of initial rule I mentioned earlier, i.e.

@RULE shapeloop-a
1 wire
2 left signal
3 right signal
4 forward signal
5 trigger right
6 idle right
7 converter
8 lead
9 edge 1
10 edge 2
11 starter
12 tempdel
13 delall
@TABLE
n_states:15
neighborhood:Moore ...

Yes, I used the same ecological experiment in the pictures above. Formally, shapeloop-ltd should have
less sensitivity to presence of boundary because ruletable may not distinguish that from empty cell,
but shapeloop-b for some configuration may do that. Why it may produce advantages for bigger ...

Comparing shapeloop-ltd and shapeloop-b for the same initial configuration on ten connected arenas, I found
that shapeloop-ltd produces less obstacles for distribution of loops:

I only should mention, that maybe redefinition I used is an overkill, because it was enough to add new state
(boundary) without any change of rule. Instead of that I defined new items to treat boundary and empty
state in the same way ... but missed some entries with empty state defined as variables ...

I think, for such competitions may be used bounded version of shapeloop rule

@RULE shapeloop-b
1 wire
2 left signal
3 right signal
4 forward signal
5 trigger right
6 idle right
7 converter
8 lead
9 edge 1
10 edge 2
11 starter
12 tempdel
13 delall
14 boundary
@TABLE
n_states:15
neighborhood:Moore ...

I suppose, it was from the "B-2 stealth" loop

x = 13, y = 15, rule = shapeloop
2$2.B2DCD$2.D3HD$2.DH.HE$2.CH.HD$2.DH.HADADA$2.AH2.4HA$2.DH5.HA$2.C7H
A$2.D2ADAD3A!


I did not guess, that particular with this detail, but "stealth" indeed looked most "creative" for me,
e.g. simplification of loops ...

Despite of mentioned technical problems, simulation on bounded grid also has some advantages due to
competition for limited space. And even shape of bounds may matter, e.g. in other thread I mentioned
400x225 bounds chosen rather due to occasional reason (16x9 display screen ratio)
http://www ...

In principle, a parent may use two arms and construct two similar shapes at once:

x = 7, y = 11, rule = shapeloop
2.2DCAD$2.A3HD$2.DH.HC$2.AH.HD$AEDH.HD$B2HJ2HB$DH.HDEA$DH.HA$CH.HD$D
3HA$DAC2D!


But basic problem after that is because each child obtained only one cell with state necessary to ...

The "movie" above shows only 10000 steps, but after that mutations are beginning and finally, after about
50 000 steps no one between initial 5 patterns survives and so it could be hardly treated
as competition. The survived loop is mutant:

x = 16, y = 14, rule = shapeloop
4$2.DCDC2DB2D$2.D7HC$2 ...

I got some problem with cut T:300,300.
Anyway, now the "competition" may be seen on YouTube for T:400,225
http://youtu.be/MGRPwuVvErs

Alexander