Thank you very much, very precise data!
That's what I did to find my alg of length 10. However my program is not efficient, I think there are ways to easily filter algs preserving corners.
I didn't know this software. It seems to be very efficient.
I thought that there is a sort of "universal" conventional definition : an edge is oriented if and only if we can put it into its original slot without F, F', B and B' moves.
But after all, it's just a convention and maybe the author of this threat uses another one.
Ok, thank you, indeed intersting that he studied also 2^2 and 4^4.
So he seems to use STM (while I'm studying in HTM).
Interesting that he find "R L2 U' F' d" as an alg with a minimal period of 2520. In HTM the highest minimal period is 1260 (reached by R' B R' U L2 for instance).
I will find a way to test the preservation of EO. ;)
Before that, some other algorithms :
0 algorithm with a height < 6 founded.
Equal to 7 moves (but take care : some of these ones are "fake" 7-moves as #3) :
Equal to 8 moves (but apparently there are a lot of fake 8-moves) :
Equal to 9...
I study the period of algorithms.
Let A be an algorithm. A is "k-periodic" if A^k (A repetead k times) is the equivalent of doing nothing.
For instance, (RUR'U') is 12-periodic, 36-periodic, and its smallest period is 6.
What I'm looking for
Here are 59 5-edge commutators finded by a program that I've created :
I don't know what they do exactly but these are 5-edges commutators.
This program searched between 1 and 7 moves. It would be easy for him to look for algorithms with longer moves if you want. Also it took about 2...