Debugging SeTT is going somewhere. The AI is partially disabled at the moment but is placing pieces on the board quite okay. Previously, it would not place any. May need to write a new calculation function. Just so happy to see some automation at long last.
But it’s almost 11pm. We had a quick and dirty dinner earlier, of roast chicken, nachos and sour cream, pickles and bread. Some banana cake for dessert. Am plotting my next coding session. Too excited to leave it overnight but the sedatives are working, putting me to sleep. More tomorrow.
…
Simplifying the code is my aim this morning and I think I can succeed. Not sure what will erupt in mindspace later. But the work is getting easier so it must be (more) correct.
Right now my idea is an anchor grabber with 4 flat conditionals, each nested with another conditional anchor grabber. This is because moves tend to happen with 3 points of reference. As for the strategy, it will be calculated from array ratios in the main AI looper. It may also be possible to have several bestMoves. How to handle that isn’t certain as is when to eat -not yet.
Took a much needed shower and am in fresh clothes. Programming is quite addictive when you’re meeting spec and passing milestones, seeing results / automation on-screen. Also nice to have a good solid computer to code on like my ThinkPad. It really feels good.
…
Night falling over PJ. My code is still unfinished as we were busy doing jobs in mindspace. Some kind of terror / blackmail / sorcery came over all of us -playthings of powerful evil witches. Hopefully I can wrap up coding after dinner. Already it’s 7pm.
…
Failed to stay awake late to code. Was waylaid by mindspace arguments and violence. At least the work is cut into manageable chunks now. Previously it was a mess of assumptions that didn’t make much sense.
Some major changes to the AI -it only cares for 5 worthwhile basic patterns and also doesn’t strategically filter for the best move. It spews out an array of choices, selected based on the strategy filter. Also streamlined the illegal move checker and randomized the first move.
Just a matter of some elbow grease and organized effort to settle the if statement coordinates which I laid out logically. It will be ready before dinner. Unfortunately, there is a power cut just at our house -the shoebox, and it's quite stuffy inside.
…
Having a simple dinner of curry sardines tonight. It’s late afternoon and I just came back from a rejuvenating run. Whacked in the if conditions and am changing the coordinates based on a diagram I made at the old place, earlier. Used reflection to simplify the process, instead of rotation.
Coordinates were changed easily. Now I’m thinking of the piece sizes returned. And the eating / impinging, if any is appropriate. Where should these be decided. Otherwise we are done!
All functions are go. SeTT moves into testing and tuning. Hopefully no horrible surprises. Have a good feeling about the code I put in over the past 2 days,
About 70% of my code is highly repetitive, checking the coordinates from multiple angles. Drawing pieces of varying sizes. Filtering moves through a precedence-staged strategy decision maker, calculating the current score. I only wrote 450 lines of clever, out of 1450. Amazing.
There are rumors SeTT may be acquired or hit the big time sometime in the near future so I have to eat that frog and keep tuning. Lots to test.
Feeling a bit sleepy. Will brush. Tune more until 12:30am.
No comments:
Post a Comment