[daip] LISTR oddity

[Michael Bietenholz]

There is something funny about LISTR's calculation of RMS with
opty=MATX: sometimes the RMS matrix is blanked when I don't think it
should be.  I think it gets the rms' values right, it just sometimes
refuses to print them.

It has something to do with weights - changing DOCALI from 100
(don't calibrate weights) to 1 cases the rms matrix to become
blanked.  I don't think this should happen, since the rms should
have a well defined value regardless of the weighting used.

On the attached data set, it works as normal with OPTY='MATX';
DOCALI=100; GAINU=1; DPARM 3 1 0; printing out matrices of first
ampl. and then rms.

However, with DOCALI=1; the rms matrix turns blank!

I did not have time to go through it in detail, but looking through
SCANUV I noticed the following which may be of relevance:

COUNT is declared float;

then it seems to get used to sum weights:

     W = VIS(IPTR+3)
     COUNT(IA1,IA2,1) = COUNT(IA1,IA2,1) + W
     COUNT(IA1,IA2,3) = COUNT(IA1,IA2,3) + W**2

So far so good, COUNT is a float sum of weights; but then COUNT is
used as if it were the integer count of visibilities in the sum

     IF ((DORMS2) .AND. (COUNT(I,J,1).GT.2)) THEN
          ... some rms calculation

Could it be that it calculates the rms based on whether the
sum of weights is >2 (rather than whether there were more than 
2 numbers to sum)?

                              michael b
-------------- next part --------------
A non-text attachment was scrubbed...
Name: testx.fits
Type: application/octet-stream
Size: 322560 bytes
Desc: 
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20071027/c270c2b4/attachment.obj>