[daip] Fringe, exhaustive baseline search gives wrong solution

[Eric Greisen]

Thomas Krichbaum wrote:
> Hi Eric,
> 
> I have run FRING on 3 other datasets. Now I do not see any problem with 
> exhaustive baseline search. If I enable exhaustive baseline search (aparm(9)=1 and set the 
> search parameter) I obtain more 'good' solutions than with ap9=0.  This is as expected and
> very good.
> 
> The main difference to the previous FRING runs of last September is that I now use 'older'
> data which where correlated with the MK4 correlator and that these older data sets did not contain
> any subarraying.
> 
> I am now wondering if a possible reason for the previous problems with the exhaustive baseline search
> may be related to either (i) DiFX (probably less likely) or (ii) to the existence of two subarrays.  
> 
> I am using FRING in version TST. Did you eventually modify FRING between Sep. 2011 and now (Dec. 12th) ?

No - I did nothing to correct FRING recently.  Your issues were too 
nebulous (from my perspective) to make debugging likely.  This may 
provide a clue.  You could take the previous data set and use UVCOP to 
separate the subarrays and then see if FRING will work.

Eric Greisen