Smallest Spaceships Supporting Specific Speeds (5s) Project

GUYTU6J · Post by **GUYTU6J** » August 5th, 2023, 5:11 am

AforAmpere wrote: ↑
August 5th, 2023, 4:19 am

GUYTU6J wrote: ↑
July 31st, 2023, 12:26 am
Well done, but you forget my modification of the script that allows for seed saving and result classifying
Sorry, I had just been looking at the ships to add and not really paying attention to too much else. I'll look it over and probably add it next update.

Okay. Feel free to alert me if anything is wrong after modification.

Also, is anyone opposed to me adding adjustables up to p1000 or so? They can't be included arbitrarily far or the database size would be infinite (unless there was some way to nicely generate them on demand), but up to p1000 seems reasonable.

Wait, is there already adjustable spaceships included in the 5S database? I prefer collecting adjustable spaceships in a separate dabatase over merging into the 5S database for two reasons: 1)interesting adjustable spaceships may have larger population than certain non-adjustable ones and thus may be excluded in the process of optimization and update; 2)for those small and competitve adjustable spaceships, having too many of them taking up smaller periods may be detrimental to diversity compared with non-adjustable ones.

In the thread I see that adjustable spaceships are usually tagged by adjustion formulae, which differs a lot from the population-displacement-period data of individual non-adjustable spaceships.

AforAmpere · Post by **AforAmpere** » August 5th, 2023, 1:58 pm

GUYTU6J wrote: ↑
August 5th, 2023, 5:11 am
Wait, is there already adjustable spaceships included in the 5S database? I prefer collecting adjustable spaceships in a separate dabatase over merging into the 5S database for two reasons: 1)interesting adjustable spaceships may have larger population than certain non-adjustable ones and thus may be excluded in the process of optimization and update; 2)for those small and competitve adjustable spaceships, having too many of them taking up smaller periods may be detrimental to diversity compared with non-adjustable ones.

In the thread I see that adjustable spaceships are usually tagged by adjustion formulae, which differs a lot from the population-displacement-period data of individual non-adjustable spaceships.

For 1), This isn't meant to replace such a thing. A collection of adjustable ships could still exist perfectly fine, but the goal of this would not be to be one. For 2), this argument has never made sense to me. This collection isn't specifically to only collect interesting ships. If anything, not minimizing them would have probably made the density of interestingness higher. It feels wrong to me to not include a ship in a collection that is supposed to collect the true minimum ship sizes only because it happens to be a member of an adjustable family. Why call it 'smallest spaceships' if it isn't actually the smallest? Also, in most cases, nothing will actually be replaced. For instance, if I added all C/n at their true minima up to p1000, more than 3/5 of them wouldn't be affected, because they already have 3 cells. I'm not sure how the last sentence applies here, each speed would be considered individually, so there'd be no need to add any new code or functionality to the database itself. One thing about the current state of known adjustable speeds too is that many of them are big, and actually increase in population with period. This would mean that most of them would likely be phased out with smaller non-adjustable ships anyway. I think I'm going to try and create a separate test database.

I think p1000 is a decent stopping point for diversity concerns as well. Many of the crazy ships are above that period anyway, so those will always stay there. It's impossible to extend the period forever without generating ships on the fly anyway.

LaundryPizza03 · Post by **LaundryPizza03** » August 7th, 2023, 4:03 am

I finally have a working B0 version of 5s. Please tell us if you find any bugs. Note that the script is based on the old version of searchRule-matchPatt2.py, not the new one with checkpointing.

sssb0.zip: 332 diagonal, 430 oblique, and 407 orthogonal ships; (51.4 KiB) Downloaded 28 times

EDIT: First results after launch.