Address feedback from Charles in a way that does not share a solution:
[Feedback from Charles]: "Another thought I have is... since at least some of this exercise is to get them thinking about complexity... shouldn't we have some tests that are actually large in scale?
If we write a method that takes an array of [row, column] coordinates for the zeros, we should be able to programmatically determine the rows and columns which need to be fully zero'd after the method is called. We could then generate input matrices of arbitrary size (10,000 x 10,000 or larger), generate random coordinates within there to zero, and automatically verify the result.
This would, I think, give a more tangible experience to how different algorithms perform."
Address feedback from Charles in a way that does not share a solution:
[Feedback from Charles]: "Another thought I have is... since at least some of this exercise is to get them thinking about complexity... shouldn't we have some tests that are actually large in scale?
If we write a method that takes an array of [row, column] coordinates for the zeros, we should be able to programmatically determine the rows and columns which need to be fully zero'd after the method is called. We could then generate input matrices of arbitrary size (10,000 x 10,000 or larger), generate random coordinates within there to zero, and automatically verify the result.
This would, I think, give a more tangible experience to how different algorithms perform."