RE: Focus indication

[Tom Droege <tdroege@veriomail.com>]

I certainly agree with Arne, a real time pipeline is the way to go.  I am 
doing my part by trying to get as many systems as possible as soon as 
possible out into the field.  Then perhaps one of you will write the 
pipeline.  Or (better) a group of you.  My approach is the monkeys and 
typewriters scheme.  ;^)  I hope "enough" is not more than 10 or 12.

Meanwhile I want to take good data this fall.  I have enough software to do 
this now.  Getting a little more information from the QBasic software would 
make it easier to line up the mounts.  I need to do two things.  Adjust the 
RA rate, and rotate the whole mount so that the RA axis points to the 
pole.  I cannot see the north pole from one of the installations, so I have 
to rotate and measure until I get it right.  Some simple measure of fwhm in 
x and y would allow me to do this.  Note that I do not have IRAF on the 
machines I am presently using.  While Arne can take an image and use IRAF 
and other tools to make such measurements, I cannot.  If an equivalent tool 
exists under Windows then I could use it.  Image Scientist will generate a 
psf from an image, but (Correct me Mike) I do not think it will give me a 
number that I can use to compare one image to the next???

Meanwhile, I will bet anyone who wants to take me on a bottle of good 
brandy that we will not have a replacement for the QBasic code by 
Christmas.  You pick the year.

Tom Droege

At 12:57 PM 8/2/01 -0700, you wrote:
>I guess it is a philosophy question, and I should
>not get deeply involved other than to add my comments
>to others who may respond with answers to your questions.
>   If the question is just how good was the night and how
>well did the equipment run that night, then my feeling is
>to use the smarts from the subsequent pipeline processing rather
>than go to considerable trouble in downfits.  For instance,
>displaying the postage stamp and allowing a user to cursor-around
>and select stars for which you display basic statistics is
>pretty simple and solves lots of problems, especially at the
>beginning of nights.  Tom wants to know if clouds are around
>later; Mike found for the Mark III that you can check for clouds
>by looking at the background and standard deviation.
>   If the question is to add real-time diagnostics for
>the acquisition code so that it can (somehow) know that
>a problem has occurred and (somehow) take some sort of action to
>solve that problem, then the course you are currently
>taking is probably as good as any.  But you really are
>starting add many of the features of a real-time pipeline
>processing system, and at some stage you should consider
>what the end result is to be and whether a fully designed
>system might get you there cleaner.
>   I'm not saying to take or reject any given approach, or whether
>to use QuickBasic or Linux.  I just have seen software systems evolve
>and in general it is better to design towards a final goal rather
>than random-walking there.
>Arne