ConwayLife.com

Posted: **December 21st, 2020, 4:19 pm**

goldenratio wrote: December 21st, 2020, 4:06 pmYep, there's a lot of overly-expensive constellations on Catagolue, which are usually caused by failed synthesis steps.
Also see this.

Isn't there a script for finding the most efficient way to synthesize things based on components? I know that the relevant constellations aren't single objects, and most of them have more than twenty bits or are oscillators instead of still-lifes, but I don't think that that matters when syntheses have already been submitted manually.

Posted: **December 28th, 2020, 5:02 pm**

MathAndCode wrote: December 26th, 2020, 8:19 pmand I submitted three-glider syntheses for the ship and tub that had two stages instead of one.

Ian07 wrote: December 28th, 2020, 1:07 pm These probably are in the database; they just haven't been chosen as the default forms of the ship and tub syntheses because Shinjuku's "prefer equally expensive syntheses with more steps" feature seems to only partially work. (from what I can tell, it seems that both versions of the synthesis must be submitted at the same time for this to actually take effect? I could be wrong about that)

So this is a little odd: it appears that the synthesis shown on xp3_ok4r2arz622611 uses a two-step version of the tub synthesis as an intermediate, whereas the tub itself (xs4_252) uses the one-step version.

MathAndCode wrote: December 21st, 2020, 4:19 pm Isn't there a script for finding the most efficient way to synthesize things based on components? I know that the relevant constellations aren't single objects, and most of them have more than twenty bits or are oscillators instead of still-lifes, but I don't think that that matters when syntheses have already been submitted manually.

Apologies for the belated response, but it's called transfer.py. However, it hasn't really been run in a while; the person who was previously handling it (Alex Greason) isn't really active anymore and no one else has taken up the task of running the script due to how long it takes to complete.

Posted: **December 28th, 2020, 5:32 pm**

Ian07 wrote: December 28th, 2020, 5:02 pm So this is a little odd: it appears that the synthesis shown on xp3_ok4r2arz622611 uses a two-step version of the tub synthesis as an intermediate, whereas the tub itself (xs4_252) uses the one-step version.

Seems like a separate submission with the two-step tub. (I think that either the pond+1G or one of the one-step tubs should be displayed on the tub page, although the one-step tub on the page seems a bit silly to me for whatever reason.)

Posted: **December 28th, 2020, 5:41 pm**

Ian07 wrote: December 28th, 2020, 5:02 pmSo this is a little odd: it appears that the synthesis shown on xp3_ok4r2arz622611 uses a two-step version of the tub synthesis as an intermediate, whereas the tub itself (xs4_252) uses the one-step version.

Using that particular two-stage version instead of the pond-based version is also somewhat odd, as the second stage takes the same amount of time, and the first stage takes much longer. One possibility (although it seems unlikely) is that the person who submitted that synthesis of xp3_ok4r2arz622611 to Catagolue actually used that particular three-glider synthesis of a tub, and that should be testable if we have an example of a staged synthesis where the first object has multiple minimal-cost syntheses, some of which take excessively long or have some other bizarre quality. (Using an unnecessarily expensive synthesis might be okay, but not necessarily.) I would suggest the block, but it turns out that of the six two-glider syntheses for a block, Catagolue uses the one that takes over twice as long to stabilize as any of the others (which is somewhat of an oddity in and of itself).

Edit: Catagolue does the same thing with the blinker. Were the Catagolue pages for the block and blinker always like this, or is this a bug with the prefer_long switch?

Posted: **December 28th, 2020, 5:53 pm**

The only objective function tracked by Catagolue is 'minimise the number of gliders'. Beyond that, it's completely arbitrary as to which synthesis ends up on the object page.

Posted: **December 28th, 2020, 5:58 pm**

calcyman wrote: December 28th, 2020, 5:53 pm The only objective function tracked by Catagolue is 'minimise the number of gliders'. Beyond that, it's completely arbitrary as to which synthesis ends up on the object page.

I thought that someone mentioned a change meant to make Catagolue prefer more stages as a tiebreaker.

Edit: I found it; it's here.

Posted: **December 30th, 2020, 2:59 am**

Apparently two isolated dots make up an xp3410.

EDIT: I will submit gmc_nxtman's 2-step 6G eater 2 synthesis to catagolue, to see if the following one-stage 6G recipe will be replaced:

Code: Select all

#C [[ GRID MAXGRIDSIZE 14 THEME Catagolue ]]
#CSYNTH xs19_4aar0rbzx121 costs 6 gliders (true).
#CLL state-numbering golly
x = 17, y = 16, rule = B3/S23
2bo$obo7bobo$b2o7b2o$11bo3$11b2o$2b3o7b2o$4bo6bo$3bo2$14b3o$14bo$
7bo7bo$7b2o$6bobo!

Posted: **December 30th, 2020, 8:16 am**

GUYTU6J wrote: December 30th, 2020, 2:59 am Apparently two isolated dots make up an xp3410.

Bizarre. The actual apgcode is xp3410_l (lowercase 'L'). It looks like all of the invalid 1yr1 objects came from hauls with apgsearch v4.972, and the ones from apgsearch v4.98 contain no such problems:

https://catagolue.hatsya.com/haul/b2-ae ... ommitted=2

I think this might be the fencepost error from ages ago which caused problems in isotropic rules whenever an unbroken horizontal line of (IIRC) >= 18 live cells occurred as a subpattern.

Posted: **December 31st, 2020, 3:14 am**

The two-loafers SVG here leaves pixels on the left edge after a few cycles.

Posted: **January 5th, 2021, 4:01 pm**

what is this rule
http://catagolue.hatsya.com/census/b3s2 ... 3232333/C1

Posted: **January 5th, 2021, 6:03 pm**

ENORMOUS_NAME wrote: January 5th, 2021, 4:01 pm what is this rule
http://catagolue.hatsya.com/census/b3s2 ... 3232333/C1

Life.

Posted: **January 5th, 2021, 7:07 pm**

but why the rulestring like that

Posted: **January 5th, 2021, 7:31 pm**

ENORMOUS_NAME wrote: January 5th, 2021, 7:07 pm but why the rulestring like that

joke/test/spam

Posted: **January 6th, 2021, 9:01 am**

ENORMOUS_NAME wrote: January 5th, 2021, 4:01 pm what is this rule
http://catagolue.hatsya.com/census/b3s2 ... 3232333/C1

Wow, a queen bee shuttle after only 10k soups?

Posted: **January 6th, 2021, 2:23 pm**

Look at the synthesis of xs21_8kiabqicz065. There are two steps that consume over forty gliders combined (including three for a LWSS), and if nothing else, why aren't they staged completely?

Posted: **January 6th, 2021, 2:26 pm**

MathAndCode wrote: January 6th, 2021, 2:23 pm Look at the synthesis of xs21_8kiabqicz065. There are two steps that consume over forty gliders combined (including three for a LWSS), and if nothing else, why aren't they staged completely?

It was submitted like that.

Posted: **January 6th, 2021, 5:08 pm**

bubblegum wrote: January 6th, 2021, 2:26 pm It was submitted like that.

I made a more efficient synthesis (based off of this soup).

Code: Select all

x = 548, y = 79, rule = B3/S23
274bo$273bo$273b3o20$5bo59bo$4bo61bo$4b3o57b3o5$72b2o$71bo2bo75b2o98b
2o$71bo2bo74bobo97bobo$2o70b2o75b2o98b2o$b2o$o$518bo$519bo$125bo391b3o
$126bo$124b3o$346b2o$346b2o5bo$352bobo101bo$352bo2bo99bobo$330bo12bo9b
2o75bo12bo11bo2bo71bo12bo$329bobo9b5o83bobo9b5o10b2o71bobo9b5o$330bo9b
o5bo83bo9bo5bo83bo9bo5bo$340bob3o2bo92bob3o2bo92bob3o2bo$234b2o105b2o
2b2o94b2o2b2o94b2o2b2o$234b2o109bo99bo99bo$346bo99bo99bo$345b2o13b3o82b
2o98b2o$248b2o110bo$249b2o110bo105b2o$248bo218bobo$467bo$254b2o$253b2o
$143b3o109bo$143bo$144bo15$283b2o$273b2o7b2o$273bobo8bo$273bo!

I have submitted it to Catagolue.

Posted: **January 9th, 2021, 10:20 am**

I might have solved the mysterious ov_s16 in 1x256X2+1. After managing to extract the soup for real this time (start with the 1x256 version, not the C1 version, since removing the $s doesn't seem to yield the equivalent 1x256 soup), I got this:

Code: Select all

x = 509, y = 1, rule = B3/S23
o5b4o3b13o2bobo2bob2o2b2obob4o3b2obo6b3obob4ob3ob4o2bo4bob4o8b2ob3ob2o
b2ob2o4b6ob3o2b2o3b2obo4bob4obo3bo2bobo8b2o2b3ob3o2bob3ob4o4b2ob2ob3o
3b5obob6ob6ob2o2b8o2b2ob6ob6ob2o2b8o2b2ob6ob6obob5o3b3ob2ob2o4b4ob3obo
2b3ob3o2b2o8bobo2bo3bob4obo4bob2o3b2o2b3ob6o4b2ob2ob2ob3ob2o8b4obo4bo
2b4ob3ob4obob3o6bob2o3b4obob2o2b2obo2bobo2b13o3b4o5bo!

There's an LWSS-LWSS bounce that results in a pair of bi-blocks:

Code: Select all

x = 159, y = 113, rule = B3/S23
2b2o151b2o$2ob2o149b2ob2o$4o151b4o$b2o153b2o106$b2o153b2o$4o151b4o$2ob
2o149b2ob2o$2b2o151b2o!

I believe that apgsearch was mistaking this as a 16-bit still life.

Code: Select all

x = 113, y = 5, rule = B3/S23
2o109b2o$2o109b2o2$2o109b2o$2o109b2o!

Another weird thing is that the same P46 showed up in xp2_1x256X2+1_Test and the original 1x256X2+1 symmetries, but one was ov_p46 and one was not. Here's the xp2_1x256X2+1_Test soup:

Code: Select all

x = 505, y = 1, rule = B3/S23
8obo2b7o2b5obo2b2obobobo2b7o3b2o2bob3o2bo2bo6b5ob2o2bob2ob3ob3o2b2o3bo
2b2obobo2bob2ob2o2b2o2b3obo3b7ob2ob2ob2ob3obobob3obo2b3ob2o2bo3b7ob2o
2bo2b4o2bob5obo2bobob2obo4b3ob5ob3ob8ob4o11b4ob8ob3ob5ob3o4bob2obobo2b
ob5obo2b4o2bo2b2ob7o3bo2b2ob3o2bob3obobob3ob2ob2ob2ob7o3bob3o2b2o2b2ob
2obo2bobob2o2bo3b2o2b3ob3ob2obo2b2ob5o6bo2bo2b3obo2b2o3b7o2bobobob2o2b
ob5o2b7o2bob8o!

And here's the 1x256X2+1 soup:

Code: Select all

x = 511, y = 1, rule = B3/S23
o2bo3bo2bob3o3b2o5b4obo5bobobob3o3bo6bo2b3ob3ob3obo2bobo6b4o2bo3b2obob
o5b4o3b2obob2o4b2o3bob3o2b2o2bo3b2o2b3o3b2o7bo2b2obo2b5obo2b3o2b3o3bob
obo3b4o2b5o2bo2b2o4bobob3ob5ob3ob8ob4ob2o2bo2b2ob4ob8ob3ob5ob3obobo4b
2o2bo2b5o2b4o3bobobo3b3o2b3o2bob5o2bob2o2bo7b2o3b3o2b2o3bo2b2o2b3obo3b
2o4b2obob2o3b4o5bobob2o3bo2b4o6bobo2bob3ob3ob3o2bo6bo3b3obobobo5bob4o
5b2o3b3obo2bo3bo2bo!

EDIT: They form from the same region, a pair of reflected 3-5-3-8-4s:

Code: Select all

x = 65, y = 1, rule = B3/S23
3ob5ob3ob8ob4o11b4ob8ob3ob5ob3o!

Posted: **January 21st, 2021, 11:25 pm**

https://catagolue.hatsya.com/object/xq6 ... y311/b3s23

plusminstrel_test
? Click it

Posted: **February 26th, 2021, 4:22 am**

What happened with b3-jknr4ity5ijk6i8s23-a4city6c7c/ikpx2_stdin ? Why can't I retrieve any soups that makes yl360_60_26422_e3691ccecad6ed5e511d701270fb5202 and yl924_2_2146_c3153ccb8b057c65a7aa6576d5295a0c — from neither the haul nor the tabulation entries?
EDIT: yl28_1_37_7373319491d9b14c3f98917f6ef45083 has the same problem as well, so I think it has nothing to do with high periods and occurence (?)

Posted: **February 26th, 2021, 11:20 am**

GUYTU6J wrote: February 26th, 2021, 4:22 am What happened with b3-jknr4ity5ijk6i8s23-a4city6c7c/ikpx2_stdin ? Why can't I retrieve any soups that makes yl360_60_26422_e3691ccecad6ed5e511d701270fb5202 and yl924_2_2146_c3153ccb8b057c65a7aa6576d5295a0c — from neither the haul nor the tabulation entries?
EDIT: yl28_1_37_7373319491d9b14c3f98917f6ef45083 has the same problem as well, so I think it has nothing to do with high periods and occurence (?)

Apologies! There was a 1920-character limit on uploaded RLEs enforced by apgsearch. I've increased it so that will record at least one occurrence if the soup is <= 5760 characters, and I'll propagate the changes to ikpx2 in the next update: https://gitlab.com/apgoucher/apgmera/-/ ... f93398470c

Posted: **March 5th, 2021, 2:11 pm**

Apparently the b3s23/C1 census says zz_LINEAR only occurred once, but there are 6 C1 occurrences in the object page.

Posted: **March 20th, 2021, 4:12 pm**

Why does Catagolue list xq4_163zzzzzzzzzzzyzym48e as requiring seven gliders?

Posted: **March 20th, 2021, 5:18 pm**

MathAndCode wrote: March 20th, 2021, 4:12 pm Why does Catagolue list xq4_163zzzzzzzzzzzyzym48e as requiring seven gliders?

Someone must have submitted an RLE containing this (possibly a Mark Niemiec mosaic of alternative syntheses for a glider) and it was misinterpreted as a synthesis for a pair of gliders:

Code: Select all

x = 42, y = 65, rule = B3/S23
35bobo$36b2o$36bo2$39bo$38bo2bo$38bo2bo$40bo49$o$b2o$2o2$7b2o$7bobo$2b
2o3bo$bobo$3bo!

Posted: **April 22nd, 2021, 2:06 pm**

Something weird is going on here, not sure if the viewer is at fault or if it's catagolue:
https://catagolue.appspot.com/object/xp ... s/g17b3s23

EDIT: definitely seems to be a Catagolue thing as the RLE is the thing that's flawed:

Code: Select all

x = 9, y = 9, rule = S23/B3/17
3.pR.qJ$2.pNpFqFqBA$.pJ3pF2qB2A$.pFRJRpNpR2A$pB2RFpRAqB2A$.V2.J2A$J$.F$2.B!

Code: Select all

x = 9, y = 9, rule = 23/3/17
3.pR.qJ$2.pNpFqFqBA$.pJ3pF2qB2A$.pFRJRpNpR2A$pB2RFpRAqB2A$.V2.J2A$J$.F$2.B!

EDIT2: okay, there seems to be State 58 cells in here for some reason:

Code: Select all

x = 9, y = 9, rule = 23/3/58
3.pR.qJ$2.pNpFqFqBA$.pJ3pF2qB2A$.pFRJRpNpR2A$pB2RFpRAqB2A$.V2.J2A$J$.F$2.B!

Code: Select all

x = 9, y = 9, rule = 23/3/59
3.pR.qJ$2.pNpFqFqBA$.pJ3pF2qB2A$.pFRJRpNpR2A$pB2RFpRAqB2A$.V2.J2A$J$.F$2.B!

ConwayLife.com

Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities

Re: Catagolue Oddities