c/8 diagonal spaceship search

[Sphenocorona]

Anybody have any good starts for making a c/8 diagonal spaceship?
Here's a quick copy of the current state of the search I'm running at the moment in WLS that has been going for a few days, which is looking for a period 4 object that moves one space horizontally and re-appears reflected across a diagonal:

Code: Select all

x = 24, y = 26, rule = B3/S23
bo$bob2obo5b2ob2o$obo3bo4b2o4bo3b3o$b2o3bo2b2obobo3bob2obo$2bo7bobo$b
2ob2obobobob2o$bo2b3o3bo2bo2bo$bo4b3o4bo2bo$bo3bo2b2o4bo$2bo10bo$3bo6b
4ob2o$2obo2b2o2b2o3bo$3b2ob2obo2b2o$obo2bob3o2bo$o3bo7b2o$3bo4bobo$2b
2o4bobobo$2bob3o$3b2o3bo$3bo$bob2o$2bobo$2bob2o$3bobo$o2bo$o!

Unfortunately there is no good definition of "partial" for diagonal and knightship sorts, which would be very helpful.

[kiho park]

Hmm.. That's a great discovery.
If we discovery about Diagonal c/8 Spaceship successfully. It's so cool story.
By the way, I think it also better to find another spaceship speeds.
So, First, finishing c/8 spaceship. then find Un-discovered Spaceships of another speed such as New Type of c/7, 3c/7 and Spaceships have speed which not in the least.





P.S. My English is too bad..

[dvgrn]

I think it also better to find another spaceship speeds.

You made the key first contribution to a 31c/240 spaceship -- doesn't that count? It's almost ready to fly now.

Unfortunately there is no good definition of "partial" for diagonal and knightship sorts, which would be very helpful.

This seems like something that is really hard to define clearly. It takes quite a bit of experience with WLS to know when it's worth continuing a search, and/or how exactly to reduce an impossibly huge search space to a manageable size.

At first glance I would say that this c/8 diagonal partial pattern is the kind of WLS result that's almost guaranteed to be disappointing... not that I have much experience with WLS myself, but I tend to worry any time a pattern stubbornly keeps getting wider and wider toward the trailing edge of the search. WLS can often figure out a way to stabilize a new cell as part of a spaceship -- but if each stabilization requires adding another unsolved cell (or two or three), then you're actually losing ground, not gaining it.

The huge size of the search space basically guarantees that WLS will eventually wander off into the weeds and start trying all combinations of ON and OFF cells in an area that will take thousands of years of CPU time to explore completely. There might actually be an answer out there, but that doesn't mean that WLS is likely to come across it before you run out of patience with the search -- or get too old to remember why you were running it...!

[kiho park]

Unfortunately there is no good definition of "partial" for diagonal and knightship sorts, which would be very helpful.

I wrote about c/7, 3c/7 diagonal "partial" spaceships.
Seeing patterns of this link, It's "almost" discovered.
viewtopic.php?f=2&t=1277

Code: Select all

#Un-Finished (1,2)c/6 thread
x = 13, y = 29, rule = B3/S23
9b2o$7bo2bo$3bo3b3o$3bo2b5obo$2b2o2bo3bobo$2b3o4b2obo$3bo5b3o$bob2o3bo
bo$bob2o3b3o$bo5bo$b2o4bob2o$2bobo2bobo$2b3ob2ob2o$2b2ob3obo$3bo6bo$5b
o$bo3b2o2b2o$b2o2bo3b2o$b4o3bo$bob2o$2o2$2obobo$2obo$3b2o4b2o$b3o$2bo
2b3ob2o$3b2o2b2o$4b3o!

From : http://plife.sourceforge.net/

You made the key first contribution to a 31c/240 spaceship -- doesn't that count? It's almost ready to fly now.

Yeah.. It seems to be finish soon. So, I thought It can be crafted.

[velcrorex]

I agree with dvgrn on this one. When searching for diagonal ships in WLS, you can set most of the cells to off and keep an open channel running diagonally from top left to bottom right. This keeps the search from getting too wide. I've searched for c/8 glide reflect ships using this method, putting in roughly a week of search time. I had some promising partial patterns, but I can't find them and haven't been keeping good records of my searches. Regardless, I think c/8 diagonal is a good place to search. Good luck!

[Sphenocorona]

I applied the suggestions above, and with a couple short searches things were already looking better. Then I switched to normal sort to try and find objects and very quickly I knew that would not work. So then I tried knightship sort and in less than a second WLS reported a partial (with the settings set to report partials every 10 columns). To me that's really promising.

Anybody else want to try running WLS or gfind to look for period 4 objects that move one space forward and flip across their backward diagonal, with a knightship sort order?

EDIT: WLS just found a partial that is a bit better, both aren't very good but they are much better than the partial at the beginning of the thread. Here's the newer, better partial:

Code: Select all

x = 15, y = 21, rule = B3/S23
2b2o$o3bo$o5b2o$bo3b3o$2b2o4bo4bo$2b2o8b2o$2b2o5bo2b2o$2bo2bo3bo2b3o$
3bo3bo4bobo$5b3o2b2o$5b2o4bobo$6b2o$5b4o$4b2o4bo$4bo3bobo$5b2o$6b2obo$
8b3o$10bo$10bo$10bo!

[kiho park]

I used to find spaceship using my notebooks. You said you find this threads using WLS search program.
As you know, Searching these threads in offline is too difficult.
So, Where can I download that program and how to execus it?

P.S. I think you might know that my English is bad...

[dvgrn]

Where can I download that program and how to [execute] it?

If you have a Windows system to run it on, you can use Jason Summers' WLS (WinLifeSearch).

If you'd rather use something that's more portable but maybe a little slower, I'd recommend JLS instead -- Karel Suhajda's Java version of the original lifesrc.

The thread includes links to a simple JLS instruction manual. Later in the thread Karel also answers some JLS usage questions with a list of advice about how to get useful results from searches. It's a good idea to spend some time working through the examples in that thread. With practice you can learn how to avoid search results that just get wider and wider without limit -- like the p8 partial result that started this thread.

P.S. I think you might know that my English is bad...

Don't worry about it. You'll get better with practice. In fact, that last P.S. had no errors in it at all!

[Sphenocorona]

Now that I have gfind, I was wondering if there is any way to get it to output some form of partial or whatnot...