I just thought I'd drop a quick note here about some fun I've been having today with satellite orbit prediction in Python. When I started mucking around with satellites, I used "predict", which was pretty good, but at some point I wanted to answer some questions which weren't easy to answer using predict. Questions like "when will AO-51 be visible from both my home in CM87 and locations in Hawaii", or "what was the radius of the circle of visibility for AO-7 compared to ISS"?
Luckily, I'm a programmer. In fact, I'm a programmer who programs for fun. So, I did a bit of research, and then coded up a version of G3RUH's "Plan 13" algorithm in Python, and then wrote some scripts to download elements from celestrak, and then a simple one to print data on the next pass of any named satellite. And, they worked pretty good. I've used them for the last year or so to do all my pass predictions. But there are still a couple of minor issues with the library. It didn't handle geosynchronous satellites very well. It implemented only the most basic of orbital models. I was never confident that the "is this satellite in eclipse" stuff working exactly right.
Luckily though, it turns out that someone else has been busy writing a more complete library: PyEphem http://rhodesmill.org/pyephem/
It's a library whose primary purpose is to calculate the positions of astronomical objects. I've used it a couple of times to (for instance) figure out the size of Mars compared to Jupiter, and found it very easy to use. But today, I realized that it had a full implementation of the SGP4 and SDP4 orbital models built in, and could be used to predict satellite passes. As a proof of concept, I hacked together a 23 line script that could print the details of upcoming ISS passes. It seems to work great, and is really quite easy to use.
You can find some of the simple example code at my blog:
http://brainwagon.org/2009/09/27/how-to-use-python-to-predict-satellite-loca...
I'll probably be porting all of my existing scripts to use this soon. In the mean time, if you have a similar task, you might look to it to solve your custom satellite prediction problems.
73 Mark K6HX