BobShemyakin wrote:I found 8-glider synthesis another quad-loaf with suppress the outer bi-loaves:
mniemiec wrote:Cute! I'm sure this could easily be adapted to a 7-glider three-loaf version.
This one is new, although the mechanism for welding loaves together has been known for a long time. Dave Buckingham synthesized the original quad-loaf from two loaves from 6 gliders. By using a loaf on one side, one can get an L-shaped tri-loaf (using your 4-glider bi-loaf or Buckingham's 3-glider one). One can also combine a loaf or bi-loaf with Dean Hickerson's tri-loaf to get yet another quad-loaf and a penta-loaf. Two tri-loafs can also be combined, by using a slightly more expensive blinker to make the butterfly predecessor from 5 gliders rather than 4 (it's also likely that a suitable 4-glider butterfly predecessor exists somewhere).
Code: Select all
x = 263, y = 178, rule = B3/S23
Extrementhusiast wrote:Also, Niemiec had made a comment that rerunning the expert system produced slightly different results. What were those different results? (In other words, which SLs should have been taken off the list, and which ones should have been put on?)
I am sorry this has taken so long. I have not had as much free time as I would have liked recently to devote to Life, so I am still catching up (for example, I have not yet assimilated your last update to the 17-bit still lifes (i.e. huge RLE with improvements to many), nor Bob Shemyakin's 6-glider still-lifes, nor even looked at any of the 18-bit still-life syntheses).
There are six irregularities in the hard 18-bit still-life list:
The following two objects had automatic solutions, so they should never have been on the list:
1) 18#146 (aka 18.817) is solvable from 33 gliders (based on 15.506 from 18)
2) 18#510 (aka 18.6770) is solvable from 33 gliders (based on 14.231 from 12)
The following object was on the list twice (I am not sure why):
3) 18#485 (aka 18.6644) is solvable from 12 gliders (based on 16.815 from 6)
4) 18#486 is the same object as 18#485
The following objects were not on the original list, but are on the list now. I am not sure why this is the case, but they can both be synthesized, so their omission from the original list is moot:
5) 18.6645 is trivially derived from 18.6644 with 2 extra gliders
6) 18.6652 is solvable from 24 gliders based on intermediary 18.6672 from 22 (based on 17#178 aka 17.1161 from 19)
The specific eater-head-to-hook converter I used above in 18.6644 and 18.6672 is fairly obvious, but was not on my converter list, so it may be new; it also solves around 0.05% of remaining unsolved larger still-lifes from 21-24 bits.
Code: Select all
x = 198, y = 116, rule = B3/S23