Forgot your password?
Puzzle Games (Games)

The Godfather of Sudoku 47

Posted by Zonk
from the brain-twisting- dept.
circletimessquare writes "The New York Times profiles 55 year old Maki Kaji who runs Nikoli, in its article Inside Japan's Puzzle Palace. Nikoli is a puzzle publisher that prides itself on 'a kind of democratization of puzzle invention. The company itself does not actually create many new puzzles — an American invented an earlier version of sudoku, for example. Instead, Nikoli provides a forum for testing and perfecting them.' Also notable is how Mr. Kaji describes how he did not get the trademark for Sudoku in the United States before it was too late. But reminiscent of a theme many Slashdotters will find familiar about intellectual property: 'In hindsight, though, he now thinks that oversight was a brilliant mistake. The fact that no one controlled sudoku's intellectual property rights let the game's popularity grow unfettered, Mr. Kaji says.' Will Nikoli be the source of the next big puzzle fad after Sudoku?"
This discussion has been archived. No new comments can be posted.

The Godfather of Sudoku

Comments Filter:
  • Godfather? (Score:5, Funny)

    by FlyByPC (841016) on Thursday March 22, 2007 @06:48PM (#18451233) Homepage
    Well, he does make a puzzle I can't refuse...
  • Nikolai (Score:5, Interesting)

    by rlbond86 (874974) on Thursday March 22, 2007 @07:10PM (#18451517)
    Nikolai makes my favorite type of puzzles, Nurikabe []. Hopefully those will pick up in the US too.
  • "Will Nikoli be the source of the next big puzzle fad after Sudoku?"

    I dunno, be he found the right place to Slashvertise, if he has it ready. (According to him, he has hundreds ready. I kind of hope he manages to get other logic puzzles popular as well.)
  • Ku? (Score:5, Funny)

    by pluther (647209) < minus pi> on Thursday March 22, 2007 @07:22PM (#18451633) Homepage
    What is ku, and why is he running it as root?
  • by Nymz (905908) on Thursday March 22, 2007 @07:31PM (#18451711) Journal
    Rather than solving Sudokus manually, I wonder how many Slashdotters have programmed their own solver, and in what languages?
    • by nebaz (453974)
      Solvers are easy, I wrote one in about 3 hours in Java. (For each square, see what permissible values it has, and try to fill things in, brute force). What is more interesting is generation. How do you know that the puzzle you've come up with has a unique solution? With a complete solution in mind, which squares can (should) you remove. That seems harder to me.
      • Re:Sudoku Solvers (Score:5, Interesting)

        by white_owl (134394) on Thursday March 22, 2007 @09:38PM (#18453155)
        Actually a generator which will generate puzzles at a given level of difficulty is not too hard if you organize the solver as a rule based system (Rule 1: first look to see if there is a square that only allows one number; Rule 2: look for pair of numbers that must exist in a pair of cells{sometimes called hidden pairs} . . . )

        After you create the rules that will solve the puzzles you categorize the rules as easy, medium or hard.

        Now if you want to create an easy puzzle you start randomly adding in (legal) numbers and each time you try to solve it using only the easy rules. The first time the solving part of the program solves the puzzle you are done. Alternatively you can start with a completed puzzle and remove a number and see if the number you removed is recoverable with the level of rules (easy, medium, hard) that you choose for the final puzzle.
    • by maxume (22995)
      Or just read the devastation of it: []
      • Norvig clearly shows how to combine constraints and search to write a solver. More interesting though is to write a solver that imitates the way people solve sudoku. Such a solver can be used for example to rate the difficulty of a puzzle, something a solver that does brute-force search can never do accurately. [To answer the parent question: Perl.]
        • by Dan Hayes (212400)
          I did this in C# and it wasn't too tricky - trying to write one for Killer sudokus though was a lot harder and I got distracted by something shiny partway through dealing with endless sets :)
          • you should pop your code over into Microsoft's Visual Studio express and get it published on Xbox Live Arcade. I'd love a Sudoku game on XBLA.

            I tried making one myself but I'm not really a programmer (it's just a hobby) and while I got a really basic brute force solver done it was ugly and not optimized very well.
        • by maxume (22995)
          I think of the puzzles as existing on a 'length to solve' or 'involvement' scale. The more a human acts like a computer, the faster they can solve them; more interesting games tend to work the other way.
    • There's a decent TCL scripted sudoku solver over here: 1 []

      Don't let the link fool you, it runs anywhere you can install the TCL shell environment.
    • by warewolfe (877477)

      I made a webpage (for practice) [] that solves sudoko puzzles. It validates in javascript and solves in php. However if the puzzle takes too long to solve the server blanks out.I'm currently writing another version in javascript to get around this. After that I'll try Perl and Java.

      If I run it as a straight php script thru the cmdline the algorithm works okay.

      Any tips for optimising server side php?

      • Re:Sudoku Solvers (Score:4, Informative)

        by Sigma 7 (266129) on Thursday March 22, 2007 @08:23PM (#18452367)

        If I run it as a straight php script thru the cmdline the algorithm works okay.

        Any tips for optimising server side php?

        It might not work, as some sudokus are extremely devilish. Instead of trying to solve it in one batch, try breaking it down into steps (e.g. have the script partly solve the system rather than having the server time the script out.)

        An example is the implementation at [] - although this is slower since it takes step-by-step very literally.
        • Re: (Score:2, Interesting)

          by warewolfe (877477)

          It's a brute force algorithm using recursive functions on a 3D array. (It seemed like a good idea at the time) but it doesn't lend itself well to stopping halfway through as it can't declare any additional numbers are right until the puzzle is solved in it's entirety.

          Unlike a logical approach which could solve Sudoku puzzle in a piecemeal fashion.

          To half solve a puzzle and give the user the chance to restart again to prevent the server timing out the page would require transferring the original state, the

          • Difficulty Rating (Score:2, Informative)

            by Nymz (905908)

            The interesting thing is, on some test puzzles that were marked difficult it took remarkably few recursions to solve while some that were marked simple took a lot longer.

            That makes sense, as ratings are based on difficulty for a human to solve.

            A difficult Sudoku might involve identifing a complex relationship in order to solve it logically, without guessing. Where as a simple Sudoku might involve only identifing simple logical relationships, but may have numerous possibilities that would increase the

    • by Chrax (782154)
      A few months back, I got a working first shot in Python. I tested it with puzzles from, and it will solve up to medium, but tends to crap out on a hard or expert one. I haven't had another of that sort of insomniatic nights since, so I haven't figured out exactly why it sometimes doesn't work.
    • I was addicted to them for a few months, and eventually wrote a solver in PHP []. Haven't done one since. However, the best solver I've seen solves them with regexes []. Regex ninjas scare the hell out of me.
      • holy freekin crap. I have a hard enough time warping my head around that stuff for basic mod rewrites in apache. ... scary ninjas indeed.
    • Re:Sudoku Solvers (Score:4, Interesting)

      by Bronster (13157) <> on Thursday March 22, 2007 @11:10PM (#18453991) Homepage
      Yeah, I wrote one in Perl. Also one for Kakuro.

      My Perl sudoku solver was probably more complete - it even does set analysis: [A, B, C] are blank, values [2, 5, 7] are available, but A can only be [5, 7], B can only be [5, 7] so C must be 2.

      It reverts to brute force only where its algorithms can't find any possible moves. I've only found that to be the case for puzzles with more than one legal solution (crappy cheap books).

      Kakuro is more interesting - first the format is a cow to transcribe to a data file. I'm using "X Y DIRN LEN SUM" as the format. Output is a HTML file because plaintext doesn't quite read easily (my sudoku solver just outputs ASCII art).

      The solver itself actually turned out to be very easy to write, and I get the feeling that it's not even complete yet (there are things I can do by hand that it doesn't attempt) and yet every puzzle I've thrown at it solves fine. It doesn't have a brute-force mode. I suspect this may be because either the kakuro book I'm using isn't very advanced, or it's just not so popular yet that people have crafted difficult puzzles.

      What I'd really love to see is a standard data file format for describing these puzzles so I can share them and/or use other peoples' puzzles as input to my scripts. But I don't have THAT much free time, really... ;)
      • What I'd really love to see is a standard data file format for describing these puzzles so I can share them and/or use other peoples' puzzles as input to my scripts. But I don't have THAT much free time, really... ;)

        My solver uses a simple XML format, so that it could easily grow elements for any future features, without obsoleting older sudoku collections. But I think I prefer doing the pen-and-paper format of daily papers. :)

        • by Bronster (13157)
          Mine uses plain text, 9 rows in which _ is an empty square, a number is a number, everything else is formatting gunk. Easy enough to write by hand, looks readable

          _ _ 1 | 3 4 5
          _ 2 3 | _ _ 7
          7 _ _ | 9 _ 8 ...

          But yeah, harder to do that sort of thing for some other puzzles.
      • I wrote a puzzle solver too, but instead of having it brute force a solution, it stops at an impasse.
        It is THOSE puzzles that are interesting. Is there something I can learn by studying the pattern? sometimes I figure out a new trick to add to my bag of sudoku solving techniques. Many times I have to resort to a parallel solution, i.e. if a box with two possibilities, if I set it as one or the other and proceed to solve, which boxes then resolve to the same number, or which boxes have the same number eli
        • by Bronster (13157)
          That's fine if your puzzle source is guaranteed to give one possible solution, as opposed to multiple possible solutions in which case you're wasting your time on an inferior sudoku source.

          That said - if _my_ solver couldn't solve past that point, I would figure out why not and encode the logic I used into it. The challenge is understanding the puzzle well enough to describe it to a computer.

          By the way - you might realise that you're describing brute-forcing (Many times I have to resort to a parallel solut
    • Prolog has got to be the way forward here. With the Eclipse constraint libraries I programmed a solver in about 10 lines. All I had to do was specify the constraints on rows (all_different and sum_to(45)) columns and squares.
  • I tried the example problems on the webpage. I had to use MSPaint, as the interface given does not support "pencil" marks and automatically assume that you are directly building the finished product.
  • by camperdave (969942) on Thursday March 22, 2007 @10:14PM (#18453491) Journal
    The following is a Binary Sudoku with a "difficult" rating:

    Good Luck!
    • by Bronster (13157)
      Has no solution in which all blocks contain all possible values. You could do one with 2x2 (1234), but that's pretty trivial.
      • Re: (Score:1, Informative)

        by Anonymous Coward
        Unless I'm reading it wrong, it is 2x2, and that's the joke. Vertical and horizontal lines delineate the cell borders.
        • Unless I'm reading it wrong, it is 2x2, and that's the joke. Vertical and horizontal lines delineate the cell borders.

          The GP is correct... each "cell" would only contain a 1 or a 0... when to be a Sudoku puzzle each cell would need to contain a 1 AND a 0.

          A good rule of thumb is that you cannot have a sodoku puzzle unless you can take the square root of the number of elements in a cell and come out with a whole number: 2x2 (1-4) is the smallest valid possibility (though you could argue a 1 cell Sodoku i

  • (Score:3, Insightful)

    by aclayton (682343) on Friday March 23, 2007 @02:22AM (#18455155)
    The article only mentions Nikoli's book-related site [], which is limited to quick introductions to a few of the puzzle types and a way to order their books. The books are beautiful, with top-notch puzzles printed on high-quality paper, but the international shipping rates make it a somewhat expensive leap of faith for someone new to these puzzles.

    A better choice is Nikoli's online puzzle site [], which is linked from their site mentioned above, but it might be easy to miss as it's not featured very well in that uncharacteristically ugly site design. There they publish 3-4 puzzles per day out of 7 different types, sizes, and varying levels of difficulty, for a monthly subscription of around $4. Nonsubscribers can play a few of each type to try out the solving interfaces, which use Flash and are far better than the ones in the NYT article. They support pencil markings, "try" modes, unlimited undo steps, partial checking, etc. What's not obvious from the nonsubscriber page is that it saves your solving histories and times for each puzzle, and you can see how you stack up to all the other subscribers. Also, you can replay your solve (with speedup/step options) to watch how you did it, or more interestingly, watch how other people did it, to expand your arsenal of techniques. You can write comments for each puzzle and read through others', but unfortunately for me most comments are in Japanese, I'm hoping the gaijin population there continues to grow..

    The variety of constructing styles and solving methods possible in puzzles like Slitherlink, Nurikabe, and Heyawake is far greater than in Sudoku, where it's often impossible to discern whether a puzzle was human-generated (i.e. Nikoli) or computer-generated (i.e. almost everything in newspapers and those bookshelves stuffed with books crapped out by every publisher trying to cash in). So for those who may have tired of relatively repetitive Sudoku solving, I recommend giving these other more elegant and interesting puzzle types a spin.

    I'm not affiliated with Nikoli, just an addicted subscriber for several years, and purchaser of many of their books. Trying to spread the word, as on one hand I'm happy to see logic puzzles like Sudoku increasing in popularity, but on the other hand I worry about high-quality puzzles getting buried by all the cash-in crud..
  • Howard Garns did, he published in magazines, it was called Number puzzle. The earliest found copies are 1976, almost 9 years BEFORE nikoli even published a "sudoku puzzle"

    The difference between number puzzle and sudoku is.... the name, it's 100 percent the same, same rules, same game.

    It's doubtful that "kakuro" is even Nikoli's either, that was another invention first found in american culture in the same magazines as number puzzle. That was called Cross Sums.

    But then again why am I expecting a massive ne
    • by The-Bus (138060)
      Did you read the article?

      "The company itself does not actually create many new puzzles -- an American invented an earlier version of sudoku, for example."
      • by kinglink (195330)
        Funny the article also claims Nikoli to be the godfather of sudoku? They are more like the taxi driver.

        "Try your hand at three different puzzles from Nikoli, a company run by the self-proclaimed godfather of sudoku,"

        Why is it that Garms isn't mentioned at all in the article. It's not hard to mention the founder of these puzzles that Nikoli appropriated. The obvious point is it's shabby journalism to call them the godfather and not clairify what they mean. An American invented the puzzle? So does that me
        • Re: (Score:3, Insightful)

          by Alzheimers (467217)
          James Brown didn't invent soul, either.
        • by The-Bus (138060)
          The New York Times did not proclaim him the godfather of sudoku. Maki Kaji proclaimed himself the godfather; Nikoli is not a person, it's the company. And godfather doesn't necessarily mean creator or originator. It means sponsor, or someone who brought something to wide acclaim and recognition. Kaji has the right to give himself that title if he wants, although it belongs to Wayne Gould just as much (who brought Sudoku to the UK).

          That you're making a tempest out of a teacup is an understatement.
  • Like most people who became enamored with Sudoku, I've grown weary of its overexposure. I don't know the exact point at which Sudoku became completely played-out, but Spongebob Squarepants Sudoku [] would probably be well past that point.

    Sudoku was fun, but the majority of the fun with any of these puzzles for me was figuring out solution methods for myself. Sudoku is now so heavily documented as to be trivial to solve, even at the highest difficulty levels (Especially if you have the patience to try Bowma

A LISP programmer knows the value of everything, but the cost of nothing. -- Alan Perlis