Summary of dicussion during chop-nod analysis telecon of Sept. 12, 07

Martin, Lero, Mike, John, Giles

_________________________________________________________________

Pointing corrections for M82:

As discussed earlier, the two options are to fit Gaussians to the I-maps
from sharpinteg or to just trust Darren's modeling technique.  John thinks
it will be straightforward for Darren to teach Lero how to make pointing
models.  The question is whether this will be accurate enough for M82
pointing corrections.  Note that sometimes Darren has done interpolations
in addition to the modelling but this is not viable for M82.  Does this
mean that for some runs this method will fail?  We need advice from
Darren.  

The Gaussian fit option, if we choose it, might be most easily done by
incorporating routines that Larry developed last year for M82 pointing
corrections.

Reduced-chi square computation:

Giles posted a document to the analysis logbook outlining a method to
compute the reduced chi squared using separate routines that act on the
sharp-combine output.  This idea had been discussed in earlier telecons.
We agreed that this seems straightforward, though John noted that there is
some ambiguity in this procedure for the case of bright sources.  
Specifically, its not clear, for bright sources, whether the gain
correction should be used or not.

Giles noted that the same routines might also be usable for data "editing"
( = outlyer rejection).  The idea is to run sharp-combine on the results
of a single file, and then compare this with the sharp-combine result for
the whole data set, to get a reduced-chi-square for that single file.

John suggested, and we agreed, that in the long run these routines should
be incorporated into sharp-combine, so that the reduced-chi-squared is
computed by sharp-combine for every data set.  One advantage of doing it
this way is that, when the reduced-chi-squared is used for data "editing",
then individual pixels in a data file can be rejected instead of having to
make a decision on a file-by-file basis.
_____________________________________________________________