Hi Stefan,

The new version does have (I47)... but the problem you describe is
a common one, with new observatory codes being added all the time.
It's probably the most frequently asked question about the program.
Which is why the new version, upon encountering an unknown observatory
code, gives the following more helpful error message:

----------- Start error message ---------
Didn't find observer XXX
Observation(s) will be excluded and treated as geocentric. You
can fix this by downloading the current list of MPC stations at

http://www.minorplanetcenter.net/iau/lists/ObsCodes.html

and saving it to the folder in which Find_Orb runs.
----------- End error message ---------

Incidentally, I've fixed the Monte Carlo issue; I'm just going through
some tests to make sure the fix didn't cause something else to break.
(Seems unlikely, but there are a few things I do want to check.)

I can highly recommend this update. It has a lot of bug fixes relative
to the previous version, plus some good new features (details under the
"revision history" at http://www.projectpluto.com/find_orb.htm ). The
statistical ranging function is more reliable than before. Linking widely
separated arcs is much more stable. There are a lot of new ephemeris
options. If you're using observations in .rwo format, you pretty much
_have_ to upgrade; the format changed a while ago.

By posting this update, I'm hoping to get all the improvements made
in the last year or so out into the world. Then, I'll start in on some
new projects. In particular:

-- I think I can improve automatic orbit determination to the point where
you can just load up observations and Find_Orb will (almost always) figure
out the orbit connecting all of them. This hasn't been a big priority,
but I've a few projects in mind for which fully automatic determination
would be a big help.

-- A little while ago, I figured out a simple way to handle uncertainties
in time. With some NEOs, it's common to receive data where the positions
appear to be solid, but the times are quite poor. In some cases, you can
say that "the following N observations have a consistent, but unknown,
error in timing"; in others, you can say that "these observations all
have good RA/decs, but time precision is absolutely awful." I think
both situations can be handled in a way that won't be too painful; with
luck, I'll have some time to implement this.

-- Uncertainties (either Monte Carlo or statistical ranging-based or
covariance-based) in orbital elements should be displayed. A little
trickier would be showing uncertainties in ephemeris positions, probably
using scatter-plots similar to those already in place on NEOCP.

-- Bill