ConwayLife.com

Congratulations to Velcrorex on discovering the Loafer! I thought this search might be worthwhile spending some CPU time on but wouldn't have dreamed of anything being found so quickly. Its fragility means that there aren't any really interesting non-destructive interactions with standard spaceships...

velcrorex wrote: Paul Tooke may have some better partial patterns and has likely put some serious search time in on this problem. According to my records that is incorrect on both counts. The only searches I did for c/7 orthogonal spaceships were a few narrow width gfind searches back in 2000. That ...

Velcrorex wrote: The ship I posted is c/4 orthogonal. Indeed! Many apologies. :oops: I confused your c/4 spaceship with a c/5 spaceship which I had (re)discovered earlier - the one that you subsequently posted in your "essential" collection. Senility is a terrible thing! :) Has anyone ever found a c...

Velcrorex wrote: Found a c/4 orthogonal ship. First of this speed in LFoD, I think. It probably would be if that was it's speed. It's more likely to be the first c/5 orthogonal spaceship! Here is a c/4 diagonal spaceship: x = 18, y = 19, rule = B2/S0 $11bo$10b3o2$8bo$8bo2bo$8b3obo$10b2o$9bobo$3b3o4b...

Velcrorex wrote: There had been an extensive pattern collection at this site http://mysite.verizon.net/live_free/ . Does anyone know if this was saved anywhere else? The Internet Wayback Machine has a snapshot of this site from March 2008: http://web.archive.org/web/20080306010610/http://mysite.veri...

EricG wrote: Unfortunately, there is an opportunity for confusion over the lettering scheme. The notation on the above webpage appears to be the original version, the one referred to in the JustFriends spaceships file, but it doesn't appear to be used anywhere else. In fact, this scheme was modified...

Mostly to EricG: You *have* been busy! It's nice to see those scripts. I had produced a simple adaption of Tim Hutton's C++ program to convert Alan Hensel's extended rule notation into a Golly rule table, but I suspect most folks would feel more comfortable with a Python script. I am still surprised...

Alegend wrote: You look for Hamiltonian paths through the graph Yes! You are obviously on the road to Damascus (not an attractive destination under present-day circumstances) but I would describe the algorithm differently. (Or perhaps I misunderstood your description.) gfind isn't looking for a Hami...

Strictly speaking this wasn't found by accident - I just wasn't expecting to find anything at all. Whilst investigating the possibility of generalising gfind to search in isotropic non-totalistic two-state rules I found that I had to consider the possibility of birth with just one neighbour. In tota...

EricG wrote: hexagonal rules can be searched! :o I'm astounded! As you're probably aware, gfind works by using five previously discovered rows a,b,c,d,e to find two new rows x & y (of which, only y is used) such that, according to the CA rule: cdy -> e & abx -> y (It also uses additional rows to pro...

Alegend wrote: The MINDEEP patch for me seems to slow it down a bit What width were you searching at? Lowering the deepening increment can slow things down because it makes the deepening round less efficient, which means that there are more of them. Therefore, although the duration of each individua...

EricG wrote: I'm bewildered - thus far at least - by how terminal() works. It is a bit cryptic in parts. The ostensible objective of terminal() of course is to check whether sufficient empty rows have been added to have found the trailing edge of a spaceship. Once the search has found 2*period empty...

William Leonard wrote: I found a very weird pattern I don't think that it is so weird: it's an 18c/72 glide-symmetric puffer. It was first discovered by Harold McIntosh in March 1994. chineseman wrote (of a 52-engine stabilization): it is so big.why not try a smaller one? Quite. Dean Hickerson did t...

Velcrorex: Your code replaces the first row generated from the root node with INITROW - but doesn't replace all of them (and generally there are going to be a lot). It will therefore have many siblings. Indeed it will be the first of the 'level 1' nodes to be searched, but it won't be the only one a...

EricG wrote: Thank you very much for going to so much work! Thanks for that. :D It's nice to have encouragement. and: I'm naively surprised that terminal() works for B2/S12 in a B2-a/S12 JustFriends search It doesn't even try to! As I said in the third paragraph of my earlier post, the changes I mad...

In the topic entitled A new GoL searching program " EricG enquired about a modified version of gfind that I used to search for spaceships in David Bell's Just Friends rule and speculated about searching in other non-totalistic 2-state rules. It is possible to search in these rules, but I only did so...

Adapting gfind Part 1 - Parameterizing the Minimum Deepening Increment. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In this posting I will describe a way to modify gfind to add an additional parameter to specify what I call the minimum deepening amount. This is the delta (<-- in...

Some time back, I started a topic Adapting Gencols in which I described some modifications that I had made to Paul Callahan's gencols program and invited similar contributions. Sadly, to date no-one has taken up this invitation :( , but discussion in the topic " A new GoL searching program " and a b...

Hmm. Me and my big mouth. I should have known that this would happen. When I mentioned the potential speed-up, all I really had in mind was producing a modified version of gfind which allowed the specification of the minimum deepening increment. Explaining all of the modifications that I've made to ...

Blinkerer wrote: I think it`s a rake Most people would call it a puffer. I suspect that you may be new to Life. Have you taken a look at Stephen Silver's Life Lexicon ? The entry for rakes explains how rakes are different from puffers. ... and: the block-laying switch engine makes zigzag of blocks.T...

velcrorex: You added what I was going to reply with whilst I was logging in! Alegend: I strongly support velcrorex's answer. As I said, these kind of searches are hard and therefore need clever algorithms. To improve on the state-of-the-art you need to understand what has preceded you. Sorry, but TA...

If you're doing c/8 and c/9 searches then the only programs I know of which are likely to find something before hell freezes over are gfind and afind , the latter being Eugene Langvagens gfind look-alike. What is it about gfind that you find too complicated to understand? The (occasionally cryptic) ...

Rake? It looks to me like a 4-switch engine block puffer. Two block laying switch engines could do the same thing.

Thanks Hartmut, that's cleared up my confusion. I'm glad you like the pattern, but it's your design not mine! I just implemented the ideas in your pattern using the tools it contained. I especially liked that pseudo-3c/7 rake. I was also struck by this reaction which occurs after the deletion of the...

Ok, I've had a week to puzzle over this and I'm still a bit confused. This pattern (suitably modified) can be be completed to contain two copies of a c/2 growing rake/puffer flotilla which cleanly stabilize the 3c/14 (not 3c/7!) diagonal pi-wave. I have a demonstration of this in the file below. But...