[Tutor] fake defrag revisited

[R. Alan Monroe]

>> achieve the cosmetic randomness, until I realized the real problem is
>> magically determining the correct sequence in which to perform the
>> moves without ruining a future move inadverently.
>>
>> If I move 0-to-1 first, I've now ruined the future 1-to-22 which ought
>> to have taken place in advance.

> For a deterministic algorithm, simply sort that list of tuples, based on
> the second item.

Thanks, although I'm not seeing how that helps:
[(63, 63), (62, 62), (61, 61), (60, 60), (59, 59), (58, 58), (57, 57),
(52, 56), (49, 55), (47, 54), (46, 53), (44, 52), (43, 51), ( 42, 50),
(41, 49), (40, 48), (39, 47), (38, 46), (36, 45), (34, 44), (33, 43),
(32, 42), (31, 41), (30, 40), (28, 39), (27, 38), (26 , 37), (25, 36),
(23, 35), (22, 34), (18, 33), (16, 32), (15, 31), (14, 30), (13, 29),
(11, 28), (10, 27), (9, 26), (7, 25), (6, 24) , (3, 23), (1, 22), (48,
21), (37, 20), (4, 19), (12, 18), (17, 17), (53, 16), (20, 15), (2,
14), (54, 13), (21, 12), (29, 11), (24, 10), (55, 9), (51, 8), (50,
7), (8, 6), (35, 5), (5, 4), (19, 3), (56, 2), (0, 1), (45, 0)]
This still suffers the same ruin-a-downstream-move problem when you
see 35-to-5 ruining the future move 5-to-4.

Alan