[daip] Re: Changes to FXPOL

[Ketan Desai]

> Chris Flatters wrote:
> > 
> > Tony Beasley <tbeasley at cv3.cv.nrao.edu> writes:
> > 
> > >
> > > Chris, surely a simple relabelling does not need to make another copy of the
> > > data? I am assuming the disk space is only needed in complex situations where
> > > data has to be moved around. pls. advise. Tony
> > 
> > It has to move data around in all cases except the one where all odd-numbered
> > bands are RCP and all even-numbered bands are LCP. The AIPS calibration
> > routines assume that RR precedes LL and will fail if it does not.

Surely Chris you only look for extra disk space if you find you must move data.
So, if most of the time, you don't need to move data, you don't need extra 
disk space.

>From the user point of view, if you run FXPOL and it says, "Hey, I need to 
move data
and there isn't enough disk space, make some room and try again." that ought to
acceptable.  

All the extra complexity of the code and the extra disk space requirements
ought to be shadowed by this check.  The need for the extra code is clear, its 
just
that it shouldnt annoy the typical user to satisfy the atypical user.

> > 
> >         Chris Flatters
> >         cflatter at nrao.edu
> 
> 
> okay, but my suspicion is that for the VLBA this covers nearly all cases;
> adding significant complexity for Euro or space VLBI cases doesn't seem like the
> way to go. Opinions?
> 
> TB


-- ketan
M.O.T.U.