So, I've been implementing this fairly complicated algorithm for the last several days. It's a port of old legacy code, and the systems engineers signed off on the details, so life should be good. But today, I finally got enough completed that I could run it on the real system. Oops, doesn't work at all. In fact, it can't work and can't be made to work. We are trying to fit a simple curve, but 95%+ of the "y" values are 0. Often, all of them will be. And no, the curve can't be a horizontal line. I really wish one of us had reallized this before I put all the effort in.
I feel like I just beat a kitten to death... with a bag of puppies.
I'd love to, but for some reason, they haven't provided me with my own multi-million dollar test aircraft to check my code against reality. All I get is an couple hours a week on a partial system that everyone knowlegable agrees doesn't work at all like the real thing.
What's so not funny is that I have used ruby on rails and some stuff that's Web 2.0 to solve problems and it's worked. In a few years, if they keep moving in the right direction, it might really not be funny.
Jason wrote:What's so not funny is that I have used ruby on rails and some stuff that's Web 2.0 to solve problems and it's worked. In a few years, if they keep moving in the right direction, it might really not be funny.
George wrote:So, I've been implementing this fairly complicated algorithm for the last several days. It's a port of old legacy code, and the systems engineers signed off on the details, so life should be good. But today, I finally got enough completed that I could run it on the real system. Oops, doesn't work at all. In fact, it can't work and can't be made to work. We are trying to fit a simple curve, but 95%+ of the "y" values are 0. Often, all of them will be. And no, the curve can't be a horizontal line. I really wish one of us had reallized this before I put all the effort in.
So, how is it that you can't output values describing the curve that are as close to a horizontal line as possible? Also, are you really modeling curves or impulses? I know you want to give away info about this SUPER TOP SECRET airplane and problems encountered that need the really clever algorithms you're using to help keep it in the air.
The airplane is a Boeing 747. There's probably nothing even proprietary about it any more; it's been around for decades. I'm sure if you wrote to Boeing's PR department, they'd be happy to provide you information about it.