16.07.2024, 19:59
I would appreciate insight from experienced puzzle setters on this proposed additional constraint option.
The proposed constraint is the option to mark some cells as "don't care" or "skip cell."
By this, I mean the cell is blacked out and is filled by no value at all, or allowed to be filled with any value, even values which otherwise would break the puzzle constraints.
Logically, a "skip cell" conflicts with no constraints, and does not cause a failure in the solution whether filled or not.
One utility is to allow setters to stop adding final resolving constraints to a puzzle that otherwise has a very interesting solution path. These final constraints can often make the earlier deductions much easier than intended, spoiling an otherwise very interesting puzzle. This option would allow you to instead just marked the unresolved cells as "skip" cells.
Some may find this approach inelegant and lead to lazy puzzle setting. Others may find it useful and beneficial.
I envision a setting procedure whereby you set sections of the puzzle to "skip" temporarily, then add additional constraints one by one, paring down the "skip" regions a little at a time until done. This would allow the software solver to converge on incomplete puzzles more easily/reliably... but it might lead to puzzles which contain unresolvable conflicts.
Might it lead to some interesting new puzzles, or better versions of puzzles?
Is this a feature you would like to see added to the tools?
The proposed constraint is the option to mark some cells as "don't care" or "skip cell."
By this, I mean the cell is blacked out and is filled by no value at all, or allowed to be filled with any value, even values which otherwise would break the puzzle constraints.
Logically, a "skip cell" conflicts with no constraints, and does not cause a failure in the solution whether filled or not.
One utility is to allow setters to stop adding final resolving constraints to a puzzle that otherwise has a very interesting solution path. These final constraints can often make the earlier deductions much easier than intended, spoiling an otherwise very interesting puzzle. This option would allow you to instead just marked the unresolved cells as "skip" cells.
Some may find this approach inelegant and lead to lazy puzzle setting. Others may find it useful and beneficial.
I envision a setting procedure whereby you set sections of the puzzle to "skip" temporarily, then add additional constraints one by one, paring down the "skip" regions a little at a time until done. This would allow the software solver to converge on incomplete puzzles more easily/reliably... but it might lead to puzzles which contain unresolvable conflicts.
Might it lead to some interesting new puzzles, or better versions of puzzles?
Is this a feature you would like to see added to the tools?
