Message #3952

From: mananself@gmail.com
Subject: Re: Cooperative Solving of Large Puzzles
Date: Mon, 15 Jan 2018 02:11:15 +0000

The first puzzle that comes into my mind is the full cell-turning 600 cell. Andrey included it in MPUlt in 2011. https://groups.yahoo.com/neo/groups/4D_Cubing/conversations/topics/1745 https://groups.yahoo.com/neo/groups/4D_Cubing/conversations/topics/1745

It has 259800 stickers. Later on, Andrey created a simplified version, keeping only about 2000 stickers. In today’s MPUlt, the "600-cell puzzle" is only the simplified one. But one should still be able to create a config to run the full version.

I don’t think anyone has attempted the full 600-cell.

I think the cooperative solving idea is possible. Many people can work on isolated areas. For example people can divide piece types, and use macros that only affect their own types.

To track progress, we can have the solution and macro definition files in a git repo, assuming macro move takes one line in such files. One can branch from master, work on their piece types, send pull requests and carefully merge back to master. There should only be line insertions in the merge but no modification.

We may also have some scripts to track the percentage of solved pieces in each type, to guard us from making bad merges. In any case, we can always go back to an old commit. I used such scripts to track some of my big solves, but working alone, I never have to create branches.

Does this approach sound reasonable?

Nan