Coma, fitting and "Star"

[andrew.bennett@ns.sympatico.ca (Andrew Bennett)]

I have posted my latest on MK IV coma and fitting on
http://www3.ns.sympatico.ca/andrew.bennett/tass/coma.html
and the latest on least squares fitting vs "Star" on
a new page (which includes stuff extracted from previous
posts) on
http://www3.ns.sympatico.ca/andrew.bennett/tass/methods.html

The conclusions of the latter are:

The aperture fitting method used by "Star" has theoretically 
higher errors than those of an optimum least squares fit. This has
been demonstrated to be the case for actual MK IV images, though
the difference in precision found for magnitude errors was not as
great as the theory predicts.

The aperture fitting method used by "Star" also introduces
rather large systematic errors when used with comatic MK IV
images.

Adoption of a new analysis method involving fitting the
variations of the PSF across the image would result in a
very large increase in computation requirements. I think we
should at least consider this. Computing power gets cheaper
every day and the potential improvement in precision is substantial.

Andrew Bennett, Avondale Vineyard, Nova Scotia, Canada.