Can someone explain the Perceptron Learning Algorithm to me? Like I'm a 5-year-old, please.

I'm trying to slog through the Caltech online course called "Learning from Data." We haven't even turned in the first homework yet, and I'm lost. We're supposed to code our own PLA:

Create 100 random points above and below a random line and designate those that are above as +1 and those that are below as -1. Got it. I can do that pretty easily.

I understand initializing all of the weights to zero.

But, I get really lost when it comes to how the algorithm iterates. I know that I'm supposed to multiply the weights across the inputs to create an output for each point. I'm supposed to compare it to the desired output and then iterate. How do I calculate desired output?

To help myself, I choice a random line y=2x and then points above and below the line. (1,1,3), (1,3,7), (1,2,3), (1,4,7) which map to +1, +1, and -1, -1, respectively.

Can anyone on Reddit walk me through the actual steps that the computer would take with these data points as it iterates? It's easy enough to copy a script from the net, but I still don't quite understand what it all does, and the mathematical notations aren't as helpful as I would have hoped.

Any help would be VERY much appreciated.