LÖVE

monolifed - something I forgot to mention, I record the tapcount and display it (reverse of your left) then at the end I get the percentage of the number of active tiles divided by the tapcount to indicate a tap efficiency, which I gave as a rating. Considered including a timer and a pause to compare rankings in speed of resolution - but that could only apply to fixed loaded Nonogram puzzles - a future development I hope.

p.s also included a hidden cheat mode.

In multiple solutions mode, initial solution is discarded. Win condition is satisfying row and column numbers. Since the solution is discarded, it doesn't know the actual correct positions.

Ah, is that the default mode. That explains my observations. I need to play around with it a bit more. Thanks for that.

Hmmm , do you run a recursive solver to check for multiple solutions, or assume there are - which makes the game a little bit harder to resolve?

Bri_G wrote: ↑Thu Aug 12, 2021 11:57 am Thought it might be that. But if you look at the image I posted - all the clues on the left say resolved whereas two at the top say unresolved.

The program is right. This puzzle has one unique solution:

Yup pgimeno, the difference is with my version the rules are you need number of tiles in row to be right and position to be right for the clue colour to be triggered. So the condition in the third row down would mean the clue on that row would be white. Means you can resolve the problem more logically.

Neat demo of yours, how did you set it up? Did you just feed in a seed value?

No, that was some GIMP work (I also used GIMP to find the solution and determine that it was unique)

I don't like that approach; it gives away too much information. If there aren't many combinations, and often there aren't, you can basically try every possibility until you find one that is correct. That's not using the numbers as clues; that's using the numbers and whether the row/column is correct or not as clues. The correct row/column indicator should just be a helper to help you focus, not a clue in itself. Considering it complete when the numbers match, does not give any more information than what you had to start with, so it's fairer IMO.

Never thought of using GIMP for that. I agree that providing both tile numbers/orientation and positions makes it a straight forward logically solvable puzzle, slow to solve initially but gets easier and faster as you progress. But that's for a simple 10 by 10 grid. As you increase the dimensions the puzzle gets progressively harder and time consuming. Take 20 by 20 and you could spend a lot of time and lose a lot of hair.

I think I'll add the option of switching positional clues on/off so that you can use logic or trial and error. I am in the process of expanding the grid size at the moment.