Matthew Francey wrote: >Minor nit: technically, the distinction between "general" and >"special" perturbations is only in how the problem is solved, not the >underlying model. True, but as a practical matter simplifying assumptions are made for general perturbations so the analytical math is manageable. SGP4 uses a simple power model for the atmosphere. NPOE, a numerical integrator "special" program (available at the Oxford ftp site), uses a much more comprehensive model (user can choose between Jacchia or U.S. Standard Atmosphere). Same goes for the gravity model. NPOE models perturbations that SGP4 omits! My personal feeling is that the need for general perturbation programs will diminish. As PC's become increasingly powerful we should reach a point where the CPU can grind through the numerical integration quickly enough that we can enjoy the greater accuracy of special perturbs without having to wait all day. >It also includes 12 and 24hr resonance effects. And actually >*looking* at the FORTRAN SDP4 code is not recommended. A most >unsettling experience it can be. :-/ I remember in college having to punch cards in Fortran 77... oops, I just dated myself :) >A more "formal" approach (using, say, Encke's method) would likely >be a little slower, but may result in "smooth" state-vector >estimates. Hm. That's what I'd like to see. NPOE is a great program but is as slow as molasses because it is 8-bit and was done in QuickBASIC. A numerical integrator done as a 32-bit program under a modern compiler would be fun to play with (anybody tried the registered NPOE, supposedly 32-bit under Microsoft FORTRAN?). _______________________________________________________ Jim Varney jamesv@softcom.net Sacramento, CA