1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
|
# TODO list
This is a list of things that I would like to add or change at some point.
It's more of a personal reminder than anything else.
## Commands
### Commands that are available in nissy 1.0, but not in this version (yet):
* drcorners (solve corners after dr)
* search and improve non-optimal subsequences
* unniss (rewrite A (B) -> B' A)
* scramble [dr, corners only, edges only, htr, ...]
* save and edit algs as "variables"
* invert an alg
### More steps for `solve`
* QTM optimal solving (important: fix possible_next, which works only for HTM now)
* Block-building steps (cross, roux blocks, ...)
* Other common steps (LSE, ...)
* optimal after eo (without breaking eo)
### Improvements to currently implemented commands
* solve should re-orient first if needed and not just give up if centers are off
* solve should try up to a small bound without loading the large pruning table
* drfin for HTR scrambles should try all 3 axis and pick the best solutions;
in general every step that automatically detects orientation should do this
* for solve -v, solving in different orientation does not give meaningful info,
because I need to transform the alg as I go.
* for solve -v, print certain info like average branching value
* solve should default to +infity for -s if other bounds are given
### New features
* cleanup: translate an alg to the standard HTM moveset + reorient at the end
* configurability: add an `alias` command, run config file at startup
* configure max ram to be used (via config file and/or command line option)
* command to transform cube and alg
* command notation to list available moves
## Distribution
* README.md: cite nxopt
* Add EXAMPLES.md file
* better man page
* webapp (cgi)
## Technical stuff
### Memory management
* free pruning table after solve is done? if so, I need to add another way
of doing batch solving (I don't want to re-load the tables every time);
for example I could add the possibility of reading scrambles from file,
and execute the same solve command to every line; also improve multi-threading:
I can just solve one scramble per thread, it's better because there is no lock.
* alternative: just add a command "free" to free up memory; it is not
user friendly (who wants to manage memory manually?) but on the other hand
it will only be used by the few who have less than 4(?) Gb of ram.
* Check if memory is enough for loading pruning tables; if not, abort
* For optimal solver: choose largest that fits in memory between nxopt and light
* Remove ptable khuge
### Performance
* solve (allow_next): filter out based on base_move; only check once for each
triple of moves; how to deal with different movesets?
### Other optimal solvers
* try htr corners + edges in slice but not oriented (300Mb table);
de Bondt's trick does not work, but I can use full symmetry and
take advantage of the fact that it is a subset invariant under half-turns
(like in light optimal solver)
* Another idea: DR + cornershtr (5Gb table); same as above, de Bondt's trick
does not work but I can use half-turn trick
### Structural changes
* client/server architecture: run a server process in the background so that
multiple client processess can send it queries and get results; this would
open up the door for a web-based version or graphical clients
### Cleanup
* Remove khuge from everywhere
* sort again functions alphabetically in their files
* more stuff to load at start (or when suitable command is called) rather
than when called directly, to avoid nasty problems with threading
|
