<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
I have some question concerning the two-body propagator PROP2B. I'd
like to use it in performing an initial validation the DIVA (from
MATH77) and TOMS Algorithm 670 (from the ACM Transactions On
Mathematical Software) numerical integrators using the two-body
problem. My questions follow.<br>
<ol>
<li>How was PROP2B validated?</li>
<li>In reviewing the code I saw that PROP2B uses a modified
version of the method documented in Danby's book on celestial
mechanics. I'm also reviewed the methods developed by Goodyear
(documented in the Astronomical Journal and elaborated in a NASA
report) and by Shepperd (documented in the journal Celestial
Mechanics). The question I have here is what are the "pros" and
"cons" when each method is compared against the others?</li>
<li>What kind of numerical precision (8 digits, 10 digits, 12
digits , etc.) can I expect from PROP2B over a one day
propagation? two days? propagation out to 8 days?</li>
<li>Lastly, are there any "gotchas" I need to pay attention to
when using PROP2B to compare two-body propagated position and
velocity vectors with their counterparts from a numerical
integration? For example, does one get better numerical accuracy
by by propagating from the epoch point to each later time point
in turn, or is it better to use PROP2B to propagate from each
time point to the next? Any run-time trade-offs between these
two approaches to pay attention to?<br>
</li>
</ol>
Looking forward to hearing from you.<br>
<br>
Sam Dupree.<br>
<br>
<br>
<ol>
<ol>
</ol>
</ol>
</body>
</html>