Learning representations by back-propagating errors (1986) [pdf]

I have been looking at a lot of explanations of back-propagation online, but since my goal is just to write a program implementing something simple, this original paper actually seems a lot more straightforward than most of the other explanations I saw.

I think actually the other explanations are explaining the math in more detail than I am able to absorb. Whereas, and I may be misinterpreting it, but the equations in this paper actually look like something I can easily type into my C++ test program. So the level of detail in this one seems more realistic for my math abilities (which are close to zero).