View previous topic :: View next topic |
Author |
Message |
| arsoncupid
| Joined: 22 Nov 2005 | Posts: 10 | : | | Items |
|
Posted: Wed Nov 23, 2005 5:20 pm Post subject: new type of Sudoku setter (request) |
|
|
Is it possible to (convince someone to) create a setting program that can take as input given numbers for a cell, as well as required candidates or required lack of candidates?
The candidates mean that, through basic logics, from the starting grid, the remaining candidates in the puzzle agree with the given candidates (or required lack thereof).
Non-basic logics:
(any backtracking methods)
n-Rook/coloring
Fishyness/Xn-wing (but basic X-wing would be okay)
Nishio/Uniqueness/templating methods
or other advanced or human-difficult solve methods
It would be nearly as useful to me to have a large database of grids that is searchable for the same thing.
You can see the usefulness of this: You want to work on puzzles that require some particular logic. Or you want to work on equivalencies to non-basic logics. Or you are looking to (dis)verify a supposed new logical rule. |
|
Back to top |
|
|
| Agent Allen
| Joined: 01 Oct 2005 | Posts: 34 | : | | Items |
|
Posted: Tue Dec 06, 2005 8:24 pm Post subject: I think you're asking for something quite easy. |
|
|
This is fairly straight forwards by the sounds of it.
The simplest solver accepts a partially solved puzzle and proceeds to solve it.
I think what you're asking is for that.
So you might insist that the leading diagonal reads 123456789 or that the top right is 2,4,6 or 8 and expect all solutions to that prescription (if any).
Is java any good to you?
The package I've knocked up can find all solutions to such prescriptions.
Lummy! You can even set them up in a GUI - except I've only really handled the unique cases very well in the GUI so far. _________________ Agent A |
|
Back to top |
|
|
|