Feedback and Questions

Hi Andrew, really like your site! It's been really useful as an instructional tool for learning the more advanced techniques. If I get completely stuck on a sudoku then I'll put it in and see what I've missed. When the solver finds the next move I'll then take a step back, switch that technique of and see if anything else works.

One thing I'd like to ask is how you choose the order of the techniques in the solver? Is it based on the difficulty of the technique? How often it occurs?

I might have a new technique for you too but I'm not sure. It's one I always use and I'd kinda assumed everyone else does to. I only really thought about this after doing Weekly Extreme No 18 from sudoku.org.uk. Once I'd moved to candidate elimination one use of the technique solved the puzzle. I'm going to wait until the weekly competition is over then post something about it on one of the forums.

I've put in the sudoku with all the numbers i'd found before having to move to candidate elimination.

Just a note on this, it wasn't particularly hard to get to this stage but the solver seems to really struggle and has to use 2 URs, an APE and a Forcing Chain to get close to this stage. I don't know if there's something missing from the solver or if this is a problem inherent in solving by candidate elimation. Any thoughts on this?

I learned something new about Sudoku strategies. Your site is extremely informative and easy to use. Prior to checking your site, I spent several days trying to solve a puzzle that required some of the advanced strategies described on your site. I simply never knew the strategies were possible. Thank you.

Thank you for the jigsaw solver.

On today's jigsaw puzzle the Law of Leftovers came up twice. I hope you have some time to show an explanation.

Also, I have been having a problem when changing canditates. I use Safari. Sometimes when I am clicking, it shuts down the browser or takes me to another place ( Sudoku solver).

I could not find todays's (jan 20) entitled Legoman in your table of shapes

When coloring first shows up, all squares with 2s and 5s, the solver
removes one number when it could remove both 2s and 5s from
the commonly situated square. Why did it only remove one ? Seems
like a bug to me.
I really like your solver. When I get a *** or a **** in my daily paper, the first thing I do is load it so that I can print out a larger version and then I let the solver go to see if it will solve and also if it involves X-wings, Y-wings or coloring, because I still have trouble understanding how to look for those.
One day, there was a totally weird algorithm, something like Bernie's Bingo. This left me speechless. "Because this one isn't 4 or 7, that one is 9". I thought, "Someone's on windowpane". But I didn't take the time to read your more detailed description of the algorithm. I guess you've added a lot since I printed out the doc last spring. Its mindblowing. Keep up the good work.

Also, if possible, please use coloring to help to determine what has changed. The originally given numbers would be one color. Then the numbers that are changing in this current move another color. Then when the next move is occurring, all non-original, past changed another color. Thanks for your consideration. I have no idea how hard this would be to code.

I like your solver. I put puzzles it to check for a single solution and to get a printable version and larger puzzle. Are puzzles usually symetric with their exposed values?
Thanks for a great solver.

I read Walter Hudiburg's comments about the helper and he is right, it is possible to solve any puzzle quickly. But I use the helper to teach me what is happening and to see if I've made any mistakes. This is a great aid and in it's present form has taught me a lot. I would hate to see it changed if the change doesn't allow us to learn and see our mistakes.

I just want to write a comment to gush about how awesome I think your solver is.

>>>>>>> G U S H <<<<<<<

Noticed that when we start filling in the board all of the loaded numbers are red, but as the board is advanced by your program some of those oriiginal values are changing to blue. This is confusing if I try to keep checking which of the new values are accurate in my incomplete puzzle.

This is just a feature request for the printed puzzles. I find that having the candidate numbers arranged in a 9x9 grid more appealing visually.

I realiize that this is just personal preference, but wonder there is a way to accomplish it - maybe using style sheets and sending the puzzle down as XML

I am one of those people who is very tech savy and even own a table PC, but I still like to solve my puzzles on paper for some reason.