>Has anyone on the UT coded a neural network using loops to achieve the desired result?
>
>Sure. What's the context?

It is for map coordinates so that a next event prediction may be made based on past events. My VFP NN code will handle up to sixty input nodes bur after that, VFP complains. So I thought that if I recode it using loops to fill arrays with weighted values instead of using variables for the weights it could handle a much larger db. I have made two attempts, but I am missing something or making an error that causes a poor result. Also, coding the NN this way makes it very slow. What I need is some example code to keep me on track.