[daip] Solaris/Linux AIPS directories

[Edward Colbert]

Hi Eric,

Well, it would be equally good to have a flag in PRTTP to be able to
include the tables (like PL) when you write a file to FITS.  I
certainly don't agree that PL files are throwaway.  They contain very
usefuly information about BLC, TRC, contour levels, and Greyscale
levels of a plot, not to mention the time and effort used in making the
plots.  All data in aips except the raw data is able to be regenerated,
so that's not really the point.  The fact that I can't write PL files
to tape is the reason I have dragged by AIPS directories around with
me instead of saving them in FITS format.  The plotting package in AIPS
is very powerful for doing overlays but this is a major drawback of
using it for that.

I know the idea is to move to aips++ anyway, but I think there would be
a lot of aips users who would like to save PL files when writing to
tape or transferring from one Endian system to another.

Ed

> From egreisen at cv3.cv.nrao.edu  Tue Oct  9 17:42:06 2001
> Date: Tue, 9 Oct 2001 15:41:59 -0600
> From: Eric Greisen <egreisen at cv3.cv.nrao.edu>
> To: "Edward Colbert" <colbert at pha.jhu.edu>, daip at primate.aoc.NRAO.EDU
> Subject: Re: [daip] Solaris/Linux AIPS directories
> 
> Edward Colbert writes:
>  > Hi,
>  > 
>  > I have been using Solaris machines at work to reduce all my AIPS data and
>  > have always preserved the AIPS directories themselves when moving from
>  > one machine to another.  The reason for this instead of writing to FITS
>  > and loading them back in is because the table attachments, especially the
>  > PL files, are unaltered.  Then, if I find a paper plot I need to re-print,
>  > or if I need to know the BLC/TRC used for a plot, etc. I can easily do so.
>  > Being able to do these things are very important to me.
>  > 
>  > I now have a new Linux machine and the floats are stored in different way
>  > so the Solaris AIPS directories will not work anymore.  Do you by chance
>  > have a utility that will convert an AIPS directory from little to big Endian
>  > and vice versa?
> 
> Unfortunately, we have never written such a program.  It would be
> quite complex --- it would have to understand the formats of all files
> - and we regard backups to tape as essential in any case using FITS
> format.  We view the PL files as throwaway - they can be regenerated
> from the data.
> 
> Eric Greisen
>