debian and astronomy (was: Re: first steps (pcfitsio, pyraf and iraf))

[justinpryzby@users.sourceforge.net]

On Mon, Oct 10, 2005 at 09:04:57AM -0400, Kevin B. McCarty wrote:
> Cedric BRINER wrote:
> 
> > Hi,
> > 
> > I'm working as a sysadmin at the Geneva Observatory and we do need
> > some softwares which are not provided by the debian distribution. So
> > I thought that it will be a good idea to do this job well once and to
> > give this back to the debian community.
> > 
> > the software that I'm willing to package are: pcfitsio, pyraf, iraf
> 
> IIRC, Justin Pryzby (justinpryzby at users.sourceforge.net) has made
> unofficial packages of at least IRAF.  There have been some possible
> copyright issues with the package.  You should probably coordinate work
> on packaging IRAF with him, especially since he owns the open ITP:
Morning,

Yes, IRAF has issues; the main CL parser is copyright Stephen C.
Johnson; I believe its ripped from 1970s unix YACC.
(./iraf-2.12.2a/unix/boot/xyacc/README; the copyright listed in the
*.c files here were apparently mass-added by NOAO, inaccurately).

I don't anticipate seeing IRAF included by Debian anytime soon.
Creating packages isn't that bad, the worst part is doing some strange
things to get it to compile (and, were I to spend more time on the
iraf packages, I would probably NOT compile from source, but write a
wget wrapper for the binaries provided by upstream, since I'm not able
to get shared libraries working with iraf's present "memory
management" scheme in conflict with my kernel's MM scheme).  There is
involving relocation of ~5 conffiles to /etc/, splitting between
/usr/lib/iraf/ and /usr/share/iraf/ ... nothing terrible here.

But IMHO the iraf system is ill-designed by today's standards.  I have
just uploaded a page about this:

  http://justinpryzby.com/astro/code/iraf.html

Then there is the rumor that NOAO is/has discontinued (financial)
support of iraf.  The references I've seen are linked to from:

  http://www.owlnet.rice.edu/~marcosh/iraf/

I don't feel that this is a bad thing .. iraf is an old system, and
probably should have been heavily rewritten several times in the last
20 years.  Now, it looks like that is actually happening.

As an aside, you might (or well might not) be interested in

  http://justinpryzby.com/astro/code/libastro.html

which is a from-scratch implementation of algorithms for astronomy
which I personally have used.  It is presently capable of image
calibration, photometry, star extraction, centroiding, moving object
detection, and includes a simiilar-triangle pattern matching routine.
(Might I add that this is 100x faster than the iraf implementation,
ccxymatch?  I can't fathom why).  I'm also working on an imtool, but
this is in very early stages of development.  Anyway, source is
X11/MIT licensed, and available as tarballs in the ./code/ directory.

I don't intend to make any use of iraf ever again, not until somebody
pays me good money to do so.

I don't expect that you're going to be able to convince your
observers/data analysts to use anything but iraf.  So I suggest my
3-step Hacker's Installation Procedure for Iraf:

Download as.pcix.gen.gz, ib.lnux.x86.gz, nb.lnux.x86.gz, and
extract them as follows:

        Extract this            To here
        -------------------------------
        as.pcix.gen.gz          ./iraf/
        ib.lnux.x86.gz          ./irafbin/bin.linux/
        nb.lnux.x86.gz          ./irafbin/noao.bin.linux/

The root directory ("./") is arbitrary; for this type of installation,
I suggest /opt/iraf/ or /usr/local/iraf/; /usr/iraf/ and /iraf/ are
other possibilities (but don't fit into the standard hierarchy as
nicely).

You must edit ./iraf/unix/hlib/libc/iraf.h with the chosen
installation path and create a symlink:

        /usr/include/iraf.h => ./iraf/unix/hlib/libc/iraf.h

That's enough to run ./irafbin/bin.linux/cl.e, and probably do most
analysis but you should run the full install script to do other stuff
too (edit other files, set up networking, make symlinks in /usr/bin/,
etc.):

        ./iraf/unix/hlib/install

(This is the part that scares most people, but I save it for last, since it
doesn't seem to be necessary for most things).

Lastly, Florian Ernst kindly sponsored my sextractor package last
week; it is now waiting in the new queue.  ds9 is already available in
Debian (though it won't be for much longer if upstream doesn't provide
unobfuscated source in their v4 release).

Iraf aside, what software do your astronomers need?  And, what iraf
tasks do they use?

-- 
Clear skies,
Justin