Hi Matt,

"...this looks as though it were derived from the same software that
TPOINT was. Is this so?"

More of an "independent co-discovery". A while back, I worked out
the math of how such a routine should work. I knew of TPoint at the
time, but didn't realize that much technical information as to how it
worked was available. So I re-derived the math from scratch.

About a year after I did all this, I dug up some text Patrick Wallace
(original author of TPoint) had written, describing the theory behind
his software. There are differences (he takes into account a few effects
I hadn't really thought through), but the basic idea is the same. That's
the good news.

"...If so, how does the program work?"

Not very well, to be truthful. First off, due to some lousy initial
design
choices, it only works with encoder-based systems of the Tangent
Instruments variety (Lumicon, Ouranos, MicroGuider, etc.) Second,
the alignment steps are not especially straightforward. More like
"user-abusive", really.

Both problems ought to be fixable. And the few people who have gotten
it to work reported substantial improvements in pointing accuracy. I've
held off on playing with it because of the work on getting Guide 8 done;
I focused on getting "data" tasks done, figuring that software-only jobs
could wait a bit. So there's some hope for improvement here. (One
reason for this is a personal one: I have an Ouranos interface,
8192-step encoders and a 13.1" mirror, plus other components, and
will be building a Dob around all this. Which will also make testing
the whole thing a lot easier.)

There is some discussion of how it works (or at least how it _ought_
to work) at

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

-- Bill