Sudoku Timer

[Kalbintion]

Why not have a high-score system for this game like others? With the fastest time being the best and difficulty obviously being a factor in the high score lists. Would help see how fast you can go for ratings and help challenge oneself to go faster.

[Multiplex]

The trouble is that the difficulty rating in the Plugin is based on the number of 'givens' - this is a bit simplistic (but better than nothing) see these two puzzles with the minimum of givens (17)
http://www.newscientist.com/articleimages/mg18825312.100/2-addictive-seductive-sudoku.html from this NewScientist articlehttp://www.newscientist.com/article/mg18825312.100-addictive-seductive-sudoku.html

Some of the smarter PC based applications seem to give a rating based in the algorithm used to solve them - not the brute force one in the Plugin (again last time I looked)

Edit:
I'm not saying it is a bad idea though.

[Kalbintion]

Though, it shan't be a hindrance of how it rates. A simple range of highscore tables could be easily assembled based on the rating given. A table for say 0 to .2, .2> to .4, .4> to .6, and so on. That being similar to a high score table for "easy", "medium", "hard", etc. Just to make sure the ranges are close enough to cause an actual competition range (obviously those numbers for the range could be modified drastically and are no in no intended way meant to be the official ones in my eyes). And why not have more than the brute force method within the game? If what you say is true, and it sounds like it makes sense, then that could be another way to do high scoring. Or hell, a multitude of tables showing that with X amount of numbers shown, a certain highscore table you're eligible for (since this is what it is now, is it now? it needn't worry about the decimal numbering that the game uses as of now)


OFFTOPIC: As a note, i wish to state that the brute force method the plugin uses is awfully pathetic. Using a simple logical sequence (1 to 9 starting on the top left, then going down 3 over to the right 1 and starting the next valid sequence of 1-9 wrapping to the left on the same line... so itd be, and once more for the last row of 3 boxes
123456789
000000000
000000000
912345678
000000000
000000000
891234567

will actually cause the method to fail epically on my sansa e200 v2 and will show a PANIC message then cause the system to need to be rebooted/restarted. Same goes for a diagonal logical method... will cause the same problem...
100004070
020000508
903000060
010400007
802050000
090306000
001040700
000205080
000030609

both are easily solvable using your head if you see the pattern right away, nothing complicated. Even the methods online sites have can solve them without having to crash themselves. Its nothing like a self-destruct sequence of 1A...2B...3C*boom*)

[Multiplex]

With reference to your comment about the solver; brute force solvers are simple and could be described as 'pathetic' - they use brute force and don't have the intelligence to spot the obvious solutions, but they are easy to write and debug, and a certain amount of optimisation is possible.

Many years ago I looked at the Sudoku plugin - I prefer Killer Sudoku these days and use a pencil and paper anyway. The standard solver would stack overflow with a simple (or completely blank) grid, so I developed an alternate brute force solver.

The patch wasn't accepted and went out of date - especially as the ability to generate puzzles was added. The importance of these comments is that in order to generate a valid grid the solver has to ensure that there is one and only one solution. For some reason the solver that was in there made fairly inefficient use of the stack so when fed a puzzle with lots of potential values it would stack overflow. I think my alternative solver may have been slower (due to the techniques used to reduce stack/memory usage) and probably didn't go further than the first solution it found (easily fixed though) - both points that might make it unsuitable for incorporation in the current plugin.

It is unfair to compare your pocket music player with an on-line application running on a machine with vastly more processing power and memory.

Also for such simple grids why do you need a solver? (you state that the solutions are easily solvable in your head)

If you were just playing, as I was, please feel free to bring my patch up to date (you may have to create a new Flyspray entry as mine has been closed) - it's number 2715, or implement a better solver.

[Multiplex]

It looks like someone else as had a look at the solver - a fix is included in the latest release.
http://www.rockbox.org/tracker/task/7772