Hello all,

My thanks to all for pitching in on trying to puzzle this out.
The error was subtle (obviously, or we'd have figured it out
long ago), and it's due to something I did, not anything Microsoft
(or Clay or anyone else) did.

I was going to send Clay a "test" version with lots of debugging
code to track down exactly where things went astray. In the process
of adding that code, I spotted what I'd done wrong.

The issue is this. A while back, I added to Guide the ability
to display UCAC-3 files from the "raw" eight-GByte dataset distributed
by USNO on two DVDs:

http://www.projectpluto.com/update8e.htm#ucac3_hd_dvd

I don't think many people are making use of this capability. There
aren't a lot of UCAC-3 disks out there, and the benefit of using the
raw data appears to be marginal in most cases. (Note that this is
a controversial viewpoint, but there are definitely some weird things
going on in UCAC-3 that we all wish weren't there.)

Anyway. Clay _is_ one of those people. Maybe the only one. And
it turns out that I added code to Guide that basically says, "If you've
drawn UCAC-3 stars, you've drawn _all_ stars; don't bother looking
at GSC or Tycho-2 or other datasets." The idea was to avoid having
stars drawn twice, as false doubles. The problem is that if those
other datasets don't get drawn, the labels associated with them don't
get drawn either.

Once I noticed this, I was able to replicate the bug. _All_ star
labels, including common names, vanished if I was zoomed in to the
point where UCAC-3 was being used. (At higher levels, the star labels
came back on.)

Fortunately, there is a way around this. The UCAC-3 stars are
flagged as to whether they exist in Tycho-2. I will revise Guide to
display "UCAC-3 not in Tycho-2", and then Tycho-2. I ought to be
able to post this fix later today (it's about two lines of code, but
I need to make sure that fixing one bug doesn't breed three more).

Thanks! -- Bill