The New Sudoku Players' Forum

[evert]

What is the most commonly used software that solves Sudoku and has a nice GUI that shows solving techniques and next steps?

Reason for asking: I did a 4-star Sudoku here and got stuck
https://www.sudokuonline.nl/

Code: Select all

418..9.......5.......4..3.9..38...6..9....1....2..3.......31..5.3..257.46....4...

[SteveG48]

Hodoku is very popular.

[gulshan212]

Hello this is Gulshan Negi
Well, Sudoku Dragon is one of the most popular programs for solving Sudoku. It has a nice graphical user interface that shows how to solve the puzzle and what to do next. It is a well-liked piece of software that provides a variety of methods for solving problems, such as coloring, X-Wings, swordfish, candidate lines, single values, and pairs and triples.
Thanks

[ghfick]

I now recommend YZF_Sudoku. This software has the most up-to-date set of techniques. Have a look at the YZF_Sudoku thread on this forum. It is in the Software folder.
I give an introduction to this software on page 20 of the YZF_Sudoku thread.

[denis_berthier]

.
To answer the question asked, i.e. "most commonly used", instead of personal preferences of a Windows-only user (ghfick) or mere self promotion (guishan212 - only 2 posts on this forum), I would line with SteveG48: Hodoku.
Hodoku has the most common used resolution rules; it is written in fully portable Java; it has a nice GUI. Download here: https://sourceforge.net/projects/hodoku/
A second choice would be Sudoku Explainer, also in Java and fully portable; mainly used for computing the SE rating - the standard rating in the Sudoku community.

[Edit]: I forgot to mention that Hodoku has very detailed online definitions of all the rules is uses, with examples: https://hodoku.sourceforge.net/en/techniques.php and a very detailed online user manual: https://hodoku.sourceforge.net/en/docs.php

[denis_berthier]

I did a 4-star Sudoku here and got stuck
https://www.sudokuonline.nl/

Code: Select all

418..9.......5.......4..3.9..38...6..9....1....2..3.......31..5.3..257.46....4...

This puzzle in particular solves with very elementary techniques:

Code: Select all

Resolution state after Singles:
   +----------------------+----------------------+----------------------+
   ! 4      1      8      ! 3      67     9      ! 25     257    267    !
   ! 3      267    9      ! 127    5      2678   ! 248    12478  12678  !
   ! 257    2567   567    ! 4      167    2678   ! 3      12578  9      !
   +----------------------+----------------------+----------------------+
   ! 157    457    3      ! 8      1479   27     ! 2459   6      27     !
   ! 578    9      4567   ! 257    467    267    ! 1      234578 2378   !
   ! 1578   45678  2      ! 1579   14679  3      ! 4589   45789  78     !
   +----------------------+----------------------+----------------------+
   ! 2789   2478   47     ! 79     3      1      ! 6      289    5      !
   ! 89     3      1      ! 6      2      5      ! 7      89     4      !
   ! 6      257    57     ! 79     8      4      ! 29     1239   123    !
   +----------------------+----------------------+----------------------+
163 candidates, 905 csp-links and 905 links. Density = 6.85%

Whips[1] are a generic view of the classical intersections/interactions/pointing+claiming. This puzzle has a lot of them at the start:

Code: Select all

whip[1]: c5n9{r6 .} ==> r6c4≠9
whip[1]: r1n2{c9 .} ==> r3c8≠2, r2c7≠2, r2c8≠2, r2c9≠2
whip[1]: r1n5{c8 .} ==> r3c8≠5
whip[1]: b9n8{r8c8 .} ==> r6c8≠8, r2c8≠8, r3c8≠8, r5c8≠8
hidden-single-in-a-row ==> r3c6=8
whip[1]: r3n2{c2 .} ==> r2c2≠2
whip[1]: b8n7{r9c4 .} ==> r6c4≠7, r2c4≠7, r5c4≠7

Resolution state after Singles and whips[1]:
   +-------------------+-------------------+-------------------+
   ! 4     1     8     ! 3     67    9     ! 25    257   267   !
   ! 3     67    9     ! 12    5     267   ! 48    147   1678  !
   ! 257   2567  567   ! 4     167   8     ! 3     17    9     !
   +-------------------+-------------------+-------------------+
   ! 157   457   3     ! 8     1479  27    ! 2459  6     27    !
   ! 578   9     4567  ! 25    467   267   ! 1     23457 2378  !
   ! 1578  45678 2     ! 15    14679 3     ! 4589  4579  78    !
   +-------------------+-------------------+-------------------+
   ! 2789  2478  47    ! 79    3     1     ! 6     289   5     !
   ! 89    3     1     ! 6     2     5     ! 7     89    4     !
   ! 6     257   57    ! 79    8     4     ! 29    1239  123   !
   +-------------------+-------------------+-------------------+
144 candidates.

The end uses only Pairs, finned X-Wings and short bivalue-chains (the most elementary form of AICs:

Code: Select all

naked-pairs-in-a-row: r4{c6 c9}{n2 n7} ==> r4c7≠2, r4c5≠7, r4c2≠7, r4c1≠7
hidden-pairs-in-a-row: r9{n1 n3}{c8 c9} ==> r9c9≠2, r9c8≠9, r9c8≠2
finned-x-wing-in-rows: n6{r6 r3}{c5 c2} ==> r2c2≠6
naked-single ==> r2c2=7
whip[1]: c6n7{r5 .} ==> r5c5≠7, r6c5≠7
whip[1]: b1n6{r3c3 .} ==> r3c5≠6
biv-chain[3]: r1n6{c9 c5} - r2c6{n6 n2} - r4n2{c6 c9} ==> r1c9≠2
whip[1]: c9n2{r5 .} ==> r5c8≠2
naked-pairs-in-a-row: r1{c5 c9}{n6 n7} ==> r1c8≠7
biv-chain[3]: r7n4{c3 c2} - c2n8{r7 r6} - b4n6{r6c2 r5c3} ==> r5c3≠4
hidden-single-in-a-column ==> r7c3=4
biv-chain[3]: c2n4{r4 r6} - r6n6{c2 c5} - b5n9{r6c5 r4c5} ==> r4c5≠4
biv-chain[3]: b4n6{r6c2 r5c3} - c3n7{r5 r9} - b7n5{r9c3 r9c2} ==> r6c2≠5
z-chain[3]: r9c2{n5 n2} - c1n2{r7 r3} - c1n5{r3 .} ==> r4c2≠5
naked-single ==> r4c2=4
hidden-pairs-in-a-column: c7{n4 n8}{r2 r6} ==> r6c7≠9, r6c7≠5
biv-chain[3]: r4c7{n5 n9} - r4c5{n9 n1} - r6c4{n1 n5} ==> r6c8≠5
biv-chain[4]: r8c1{n9 n8} - c2n8{r7 r6} - r6n6{c2 c5} - r6n9{c5 c8} ==> r8c8≠9
singles ==> r8c8=8, r8c1=9
biv-chain[4]: r6c2{n6 n8} - r7c2{n8 n2} - r7c8{n2 n9} - r6n9{c8 c5} ==> r6c5≠6
singles ==> r6c2=6, r3c3=6, r7c2=8
biv-chain[3]: r5c4{n2 n5} - r5c3{n5 n7} - c6n7{r5 r4} ==> r4c6≠2
singles ==> r4c6=7, r4c9=2
biv-chain[4]: c8n5{r5 r1} - c8n2{r1 r7} - r7c1{n2 n7} - c3n7{r9 r5} ==> r5c3≠5, r5c8≠7
stte

Check where you get stuck. That will tell you the next rule you have to learn.
.

[ghfick]

For: 418..9.......5.......4..3.9..38...6..9....1....2..3.......31..5.3..257.46....4...

HoDoKu gives a default solution path of:

Code: Select all


Hidden Single: r1c4=3
Hidden Single: r2c1=3
Hidden Single: r8c4=6
Hidden Single: r7c7=6
Hidden Single: r9c5=8
Hidden Single: r2c3=9
Naked Single: r8c3=1
Locked Candidates Type 1 (Pointing): 5 in b1 => r3c8<>5
Locked Candidates Type 1 (Pointing): 7 in b8 => r256c4<>7
Locked Candidates Type 1 (Pointing): 9 in b8 => r6c4<>9
Locked Candidates Type 1 (Pointing): 8 in b9 => r2356c8<>8
Hidden Single: r3c6=8
Locked Candidates Type 1 (Pointing): 2 in b2 => r2c2789<>2
Locked Candidates Type 1 (Pointing): 2 in b1 => r3c8<>2
Naked Pair: 2,7 in r4c69 => r4c125<>7, r4c7<>2
Naked Triple: 2,8,9 in r78c8,r9c7 => r9c89<>2, r9c8<>9
Skyscraper: 6 in r2c6,r3c3 (connected by r5c36) => r2c2,r3c5<>6
Naked Single: r2c2=7
Locked Candidates Type 1 (Pointing): 7 in b2 => r56c5<>7
AIC: 4 4- r4c2 -5- r9c2 =5= r9c3 -5- r3c3 -6- r3c2 =6= r6c2 =8= r7c2 =4= r7c3 -4 => r5c3,r7c2<>4
Hidden Single: r7c3=4
Discontinuous Nice Loop: 2 r1c9 -2- r4c9 =2= r4c6 -2- r2c6 -6- r2c9 =6= r1c9 => r1c9<>2
Locked Candidates Type 2 (Claiming): 2 in c9 => r5c8<>2
Naked Pair: 6,7 in r1c59 => r1c8<>7
Discontinuous Nice Loop: 5 r4c2 -5- r9c2 =5= r9c3 =7= r5c3 =6= r6c2 =4= r4c2 => r4c2<>5
Naked Single: r4c2=4
Naked Triple: 2,5,9 in r149c7 => r6c7<>5, r6c7<>9
XY-Wing: 1/9/5 in r4c57,r6c4 => r6c8<>5
XY-Chain: 2 2- r3c1 -5- r4c1 -1- r4c5 -9- r4c7 -5- r1c7 -2- r9c7 -9- r9c4 -7- r9c3 -5- r9c2 -2 => r3c2,r7c1<>2
Hidden Single: r3c1=2
Locked Candidates Type 2 (Claiming): 5 in c1 => r5c3,r6c2<>5
XY-Wing: 6/8/7 in r5c3,r6c29 => r5c89,r6c1<>7
Locked Candidates Type 1 (Pointing): 7 in b4 => r5c6<>7
Hidden Single: r4c6=7
Naked Single: r4c9=2
XY-Chain: 1 1- r4c1 -5- r4c7 -9- r9c7 -2- r9c2 -5- r9c3 -7- r5c3 -6- r5c6 -2- r5c4 -5- r6c4 -1 => r4c5,r6c1<>1
 stte



stte is an abbreviation for 'singles to the end'

YZF_Sudoku gives almost the same default solution path:

Code: Select all


Hidden Single: 6 in b9 => r7c7=6
Hidden Single: 6 in b8 => r8c4=6
Hidden Single: 8 in b8 => r9c5=8
Hidden Single: 3 in b1 => r2c1=3
Hidden Single: 3 in b2 => r1c4=3
Hidden Single: 9 in b1 => r2c3=9
Naked Single: r8c3=1
Locked Candidates 2 (Claiming): 2 in r1 => r2c7<>2,r2c8<>2,r2c9<>2,r3c8<>2
Locked Candidates 2 (Claiming): 5 in r1 => r3c8<>5
Locked Candidates 1 (Pointing): 7 in b8 => r2c4<>7,r5c4<>7,r6c4<>7
Locked Candidates 1 (Pointing): 9 in b8 => r6c4<>9
Locked Candidates 1 (Pointing): 8 in b9 => r2c8<>8,r3c8<>8,r5c8<>8,r6c8<>8
Hidden Single: 8 in r3 => r3c6=8
Locked Candidates 1 (Pointing): 2 in b2 => r2c2<>2
Naked Pair: in r4c6,r4c9 => r4c1<>7,r4c2<>7,r4c5<>7,r4c7<>2,
Hidden Pair: 13 in r9c8,r9c9 => r9c8<>29,r9c9<>2
Skyscraper : 6 in r2c6,r3c3 connected by r5c36 => r2c2,r3c5 <> 6
Naked Single: r2c2=7
Locked Candidates 1 (Pointing): 7 in b2 => r5c5<>7,r6c5<>7
Uniqueness Test 7: 27 in r45c69; 2*biCell + 1*conjugate pairs(7c6) => r5c9 <> 2
Uniqueness Test 7: 89 in r78c18; 2*biCell + 1*conjugate pairs(8c8) => r7c1 <> 9
Hidden Single: 9 in b7 => r8c1=9
Full House: r8c8=8
Discontinuous Nice Loop: 7r4c6 = (7-2)r4c9 = (2-6)r1c9 = r1c5 - r2c6 = (6-7)r5c6 = 7r4c6 => r4c6=7
Hidden Single: 2 in r4 => r4c9=2
Naked Pair: in r1c5,r1c9 => r1c8<>7,
Discontinuous Nice Loop: 4r7c3 = (4-8)r7c2 = (8-6)r6c2 = (6-4)r5c3 = 4r7c3 => r7c3=4
Discontinuous Nice Loop: 4r4c2 = (4-6)r6c2 = r3c2 - (6=5)r3c3 - r9c3 = r9c2 - (5=4)r4c2 => r4c2=4
Hidden Pair: 48 in r2c7,r6c7 => r6c7<>59
XY-Wing: 159 in r4c5 r4c7 r6c4 => r6c8 <> 5
XY-Chain: (2=5)r3c1 - (5=1)r4c1 - (1=9)r4c5 - (9=5)r4c7 - (5=2)r1c7 - (2=9)r9c7 - (9=7)r9c4 - (7=5)r9c3 - (5=2)r9c2 => r3c2,r7c1<>2
Hidden Single: 2 in b1 => r3c1=2
Locked Candidates 2 (Claiming): 5 in c1 => r6c2<>5,r5c3<>5
W-Wing: 78 in r6c9,r7c1 connected by 8r5 => r6c1<>7
Locked Candidates 1 (Pointing): 7 in b4 => r5c8<>7,r5c9<>7
XY-Chain: (5=9)r4c7 - (9=1)r4c5 - (1=5)r6c4 - (5=2)r5c4 - (2=6)r5c6 - (6=7)r5c3 - (7=5)r9c3 - (5=2)r9c2 - (2=9)r9c7 - (9=2)r7c8 - (2=5)r1c8 => r1c7,r5c8<>5
 
stte



HoDoKu used an early notation for chains/loops while YZF_Sudoku uses the Eureka notation for chains/loops. Eureka is 'commonly used' these days.

[denis_berthier]

.
Both notations for chains are inconsistent. They deal with bivalue in rc-cell and bilocal in different ways.

[ghfick]

Regarding the GUI [ Graphical User Interface ], HoDoKu and YZF_Sudoku give very similar images of the grid with visuals of each technique used in the current solution path.
YZF_Sudoku does give all four views [ RC, RN, CN and BN ] and, using the "Color Cell" tab, you can construct your own visuals of chains.

HoDoKu was last released in 2012. YZF_Sudoku's current version was released in July of 2022.
YZF_Sudoku now contains GSP, Almost Locked Pair, Almost Locked Triple, ERI Pair, Uniqueness Loop, Bivalue Oddagon, Extended Rectangle, Firework, Broken Wing, Exocet: [ Junior, Senior, Weak ], MSLS, Triplet Oddagon along with many Dynamic Chains [ from Sudoku Explainer ] and a collection of techniques being called Memory Chains.

I should mention XSudo :

sudoku.allanbarker.com/

XSudo has a fascinating GUI. It is very distinctive. There is a 3D selection too! Many Sudoku people use parts of XSudo's terminology. XSudo separates the vast collection of AICs into S-Wings, M-Wings, H-Wings and Purple Cows. Loops are separated into various Rings. There is much more in XSudo. I believe the last release of XSudo was in 2010.

Also very interesting is Stormdoku :

forum.enjoysudoku.com/stormdoku-t32977.html

Stormdoku has a vast collection of techniques. A huge collection of wings and rings and many, many more. Stormdoku is constantly being updated. The most recent version is from August of 2022. So far, I believe that Stormdoku does not have a GUI. Stormdoku is written in Turbo Pascal and has been designed for MSWindow OS's. I do not have a machine with MSWIndows so I have not used Stormdoku myself. My programming skills are very modest but I understand that those with better skills than me could, in principle, set up Stormdoku on machines with either Mac or Linux.

I would be remiss if I did not mention Andrew Stuart's site :

www.sudokuwiki.org/

Andrew is constantly updating his site. Even as recently as January 2023. The GUI for his solvers is very fine. I understand that he has plans to add a 9x9 KenKen solver in the future. I believe his solvers will run with just about any web browser.

Also really good is Philip Beeby's site :

www.philsfolly.net.au/

Philip is also constantly updating his site. His solver has techniques not seen anywhere else, I believe. His solver has a collection of Complex Chains, a new technique called Replacement, parts of the Pattern Overlay method and a number of techniques based on Contradiction, Networks, Sets and much more. His solver runs very well with some [ but not all ] web browsers. I am not sure about this matter currently.

One more mention: Sudoku Explainer [ SE ] by Nicolas Juillerat is very good at, well, explaining:

forum.enjoysudoku.com/sudoku-explainer-t39865.html

The original version by Juillerat dates from December 2006. There are several [ many! ] versions of SE noted in the forum site. SE offers detailed explanations of each step in a solution path. The explanations are particularly helpful when one is learning the techniques [ that were available at the time ]. SE offers a massive collection of Dynamic Chains. Dynamic chains are not for the faint of heart. SE does give explanations of these Dynamic Chains. The explanations can go on for several pages.

[denis_berthier]

YZF_Sudoku now contains ... a collection of techniques being called Memory Chains..

A name that hides plagiarism, as mentioned in the yzf-sudoku thread. It's really surprising that all the other techniques get their right name, but not whips, g-whips and the like.
As you perfectly know this, I wonder why you are propagating such misinformation.

I'll also take this opportunity to recall that the four views: rc, rn, cn and bn were first introduced in my book: "The Hidden Logic of Sudoku". At least, they retain the name I gave them.

Stormdoku is written in Turbo Pascal and has been designed for MSWindow OS's. I do not have a machine with MSWIndows so I have not used Stormdoku myself. My programming skills are very modest but I understand that those with better skills than me could, in principle, set up Stormdoku on machines with either Mac or Linux.

NO. Nothing written for Windows can run on a recent Mac. Wine (which might have been a solution for old Macs) will not be ported to the recent versions of MacOS. We have also already talked about this, so I don't understand why you keep repeating it.

To be more specific, here are the compatibilities with the last versions of OSes: (from https://wiki.winehq.org/Download)

Code: Select all

Ubuntu - WineHQ binary packages for Ubuntu 18.04, 20.04, 22.04, and 22.10
Debian - WineHQ binary packages for Debian Bullseye and Bookworm
Fedora - WineHQ binary packages for Fedora 36 and 37
macOS - WineHQ binary packages for macOS 10.8 through 10.14

The current version of MacOS is 13.2.1, with lots of security additions that block any unchecked software.

[ghfick]

NO. Nothing written for Windows can run on a recent Mac.

FALSE. StrmCkr has kindly made his source code available. Someone with some Pascal experience could edit and rewrite some sections of the MSWindows specific code.
This code could then be compiled with a Mac version of Pascal or a Linux version of Pascal.
Further, at the present time, there are no maintainers of the Mac WineHQ binary packages. This could change. Are there volunteers out there?

[denis_berthier]

.
You're proving exactly the contrary of what you claim. The Windows code had to be modified for running on the Mac. This required having the source code.
As for Wine, you confirm what I said. Moreover, there will never be a full version of Wine acceptable by Apple, because it would break the security requirements of MacOS.

[ghfick]

Open source is rapidly becoming the norm. So open source code can then always, with revision, run with any OS.
You are going around in circles. Good luck with that.

[denis_berthier]

.
Stick to facts instead of personal fantasies. Wine doesn't run on MacOS and there is no prospect for it do so so in any foreseeable future. Fullstop.

[ghfick]

You are definitely not an expert on Mac matters or Wine matters. There is no way you can know that the status quo will maintain. Who knows whether someone will volunteer to make the steps forward.