Day 24

Today was heavily focused on the 4x4 implementation of my Tic Tac Toe Board. This story was really involved mostly because my existing code wasn’t following OCP well enough. I had to refactor old code just to be able to extend new implementation, so I spent a lot of time making it compliant with OCP. This mostly involved changing how my menu screen allows user selection, as well as being open to any size board and how those inputs would be processed. I found this work to be quite rewarding because I can go back on my earlier work and understand myself exactly how I could have done it better. This is practice that after a while of continually hammering down, I can hopefully follow without even thinking about it!

After my code was looking open for extension, I needed to implement a 4x4 board, the meat of this story. Getting it to work for users was fine, just a bit of fiddling with the inputs, but the AI implementation was incredibly difficult. Early on I got it working in theory, but the computation time was way too long to be acceptable. When doing an involved tree search through 9 levels of permutations, it takes a few seconds, but at 16 levels for the 4x4 board, it was more likely hours or even days.

I started researching solutions for this issue and came across two concepts, limited depth search and alpha-beta-pruning. Both were pretty hard to implement and get my AI to behave as it should. After half a day of attempting, I was not successful in getting this working, but I feel as though I’m close, tomorrow I will finish this off.