Hi folks,

Thanks for the comments on the incorrect version dates in the
updated software. I have made that mistake all too frequently, and
have just now done what I should have done about a year ago: to the
batch file that produces the update .ZIPs, I've added the lines:

echo Remember: Check the version dates in *HELPEXT.DAT!
pause

(Files such as EHELPEXT.DAT contain the version date as the final line.)
It should give me the "nudge" I need to get things right.

About asteroid occultations: prepare for a lengthy speech <grin>.

You can click on the asteroid, then click on the star, then click
on "Show Eclipse" to see the ground path. I toggled Guide to "Use MPCORB",
clicked on (752) Sulamitis, clicked on Mu Geminorum, clicked on "Show
Eclipse", and... I got essentially the result you suggested: an
occultation _just barely_ occurs in the North Atlantic.

This didn't really surprise me. MPCORB.DAT always has an epoch of
osculation near the present; right now, it's set for 13 September
2000 (JD 2451800.5). In trying to predict the asteroid position for
20 November (date of the occultation), about two months of perturbations
are ignored. That throws it off by an arcsecond or so.

Turning off MPCORB and using Guide's built-in data, which is stored
at many different epochs, got me an orbit with an epoch of osculation
of 2 November 2000. Repeating the click-on-asteroid, click-on-star,
"Show Eclipse" got me a much better path, which I screen-captured and
posted at

http://www.projectpluto.com/752.gif

It runs a bit north of the "old" path shown on page 109 of the
November _Sky & Tel_. In general, paths generated by this method
will be in the right ballpark, but that's about all.

This is a good place to repeat something I've said before: don't
use MPCORB for dates distant from the present, ever. (Not that this
is a rap against MPC. If Guide included the machinery of a full-fledged
numerical integrator, it could get decent results with MPCORB, even
for distant dates.)

The method used by Martin Federspiel, and by Jan Manek and others,
is indeed available in Guide. All that is involved is computing a
highly accurate orbit using precision astrometry. (MPCORB and ASTORB
and Guide's built-in data are not sufficiently accurate, and probably
won't be for a long time.) If you look at...

http://www.projectpluto.com/ast_occu.htm

...you'll see a slew of predictions based on this method.

There are two sticking points to this. You have to get highly
precise astrometry (Tycho-2 or ACT or UCAC based), _and_ you need
to process this into an equally precise orbit with an epoch of
osculation near that of the occultation, using software such as
Find_Orb.

I'm currently getting the "highly precise astrometry" from Ron
Stone, at the Naval Observatory in Flagstaff, Arizona, plus some
bits from Gordon Garradd and from Table Mountain. As far as I know,
these three groups/individuals are the only people producing this
high-quality astrometry. (If you have a CCD with a wide field, this
is an _excellent_ area to get into if you're looking to produce
scientifically useful data.)

I run these through Find_Orb, get the "precise orbit with an
epoch of osculation near that of the occultation", save the elements
to a file, then load those elements using Guide's "Add MPC Comets/
asteroids" function.

I started to write up a description of the details in this, and
realized that it's very tricky to describe. I'll give it another
shot in the next few days, and will announce it when it's posted.

-- Bill