Marcel Hendrix wrote: "Gerry" <gerry@jackson9000.fsnet.co.uk> writes Re: Sudoku3 [..] Anton's #1 and #4 didn't need step 3, the rest of his did. #3 does appear to be the hardest and is the worst I've seen. However it found the 'minimal' puzzle rather easy as it didn't need to go to step 3. That seems strange as it took quite a long time in your results. ...

"Gerry" <gerry@jackson9000.fsnet.co.uk> writes Re: Sudoku3 [..] Anton's #1 and #4 didn't need step 3, the rest of his did. #3 does appear to be the hardest and is the worst I've seen. However it found the 'minimal' puzzle rather easy as it didn't need to go to step 3. That seems strange as it took quite a long time in your results. Obviously, what is difficult for a human being...

Marcel Hendrix wrote: anton@mips.complang.tuwien.ac.at (Anton Ertl) writes Re: Sudoku3 [..] My program solves only puzzle1 and puzzle4 from <http://www.complang.tuwien.ac.at/anton/lvas/effizienz-aufgabe06/> completely. Either your program is better already without backtracking or these puzzles are relatively difficult. Using Robert Spykerman's code: ...

anton@mips.complang.tuwien.ac.at (Anton Ertl) writes Re: Sudoku3 [..] My program solves only puzzle1 and puzzle4 from <http://www.complang.tuwien.ac.at/anton/lvas/effizienz-aufgabe06/> completely. Either your program is better already without backtracking or these puzzles are relatively difficult. Using Robert Spykerman's code: speedthem ( 3 GHz PIVht ) 0.195 ...

From: "Gerry" <gerry@jackson9000.fsnet.co.uk> Newsgroups: comp.lang.forth Subject: Re: Sudoku3 Date: Sat, 30 Dec 2006 07:27:08 -0800 Anton Ertl wrote: [snip] Certainly Prolog, especially with finite domain constraints, is a very good language for solving Sudoku and other puzzles. However, we already had a nice fast solver in Forth by Robert Spykerman, so the aim of my ...

anton@mips.complang.tuwien.ac.at (Anton Ertl) writes Re: Sudoku3 [..] I have added the few missing ANS Forth words and put this program at http://www.complang.tuwien.ac.at/forth/programs/sudoku4.fs Let it be known that I immediately and completely lost interest in Sudoku type puzzles when a slightly adapted version of Robert Spykerman's program became able to solve all of them for ...

...less than I had originally envisioned. Also, other things intervened, so it took much longer than I had planned. Anyway, you can find the program at http://www.complang.tuwien.ac.at/forth/programs/sudoku3.fs It relies on a number of Gforth words (see the comments at the start). It is also bigger and slower than Robert Spykerman's program, and it solves only simple puzzles at the moment...

"Gerry" <gerry@jackson9000.fsnet.co.uk> writes: Anton's #1 and #4 didn't need step 3, the rest of his did. Likewise my solver only solves #1 and #4 completely. So it appears that your rules 1 and 2 are indeed the rules that I implemented. However it found the 'minimal' puzzle rather easy as it didn't need to go to step 3. Yes, my solver also solved that. - anton -- M. Anton ...

... the partial "second level" rules seems to solve all of the "hard" puzzles I find locally and at the WebSudoku page and the "Fiendish puzzle April 21,2005 Times London" from Andreas Kochenburger in the Sudoku3 thread. I also have another puzzle from my local paper that I have not been able to solve by hand. It is shown here. I'm not sure it has a single solution as I recall ...

...Adding the partial "second level" rules seems to solve all of the "hard" puzzles I find locally and at the WebSudoku page and the "Fiendish puzzle April 21,2005 Times London" from Andreas Kochenburger in the Sudoku3 thread. I also have another puzzle from my local paper that I have not been able to solve by hand. It is shown here. I'm not sure it has a single solution as I recall reading ...