iraf.net

We (Penn State) recently installed IRAF V2.10 on our network. I was under the
impression that the "old" newimred package stuff was to be incorporated into
V2.10 more or less as was (with any bug fixes, of course). Most of the work
I do is with echelle spectra, so I used the newimred.echelle package, which is
now imred.echelle. I just reduced my first batch of echelle data with the
V2.10 software, and am horrified by what I find: my image headers (after
extraction) no longer have the APNUM# fields that specify the particulars for
each order. Instead, I find the WAT* fields. While they give the same info,
I wonder why the APNUM# format was abandoned, as it was easier to read. My
real concern is that I had written several IMFORT programs that could deal
with "onedspec" and "echelle" format data. I was happy to rewrite my code to
support the echelle format, as well as onedspec, when it appeared. The format
made good sense. I am not thrilled about having to rewrite my programs yet
again for what appears to be no good reason. So, I have some questions:
1) Have I missed something in the transition to V2.10? Is the APNUM# header
format still available somehow?
2) Assuming a "no" answer to the above, what was the reason for doing away
with the "old" format?
3) On the same assumption: is there any hope of convincing you to reinstate it?For the present I think I'm just going to reinstall the newimred add-on
package...
Thanks,
Alan
welty@astro.psu.edu

Hi Alan,The NEWIMRED packages were a (very useful) half-way point between V2.9 and
V2.10. In V2.10 the capabilities of NEWIMRED (the APEXTRACT and ONEDSPEC
changes and the reduction scripts) are incorporated in the distributed
NOAO packages. The main change in the second half of the revisions was
to add support for world coordinate systems (WCS). The concept of a WCS
in IRAF is that operations within and outside a package always keep the
world coordinates (the dispersion coordinates in the case of spectroscopy)
correct. For example copying a subsection of a spectrum with IMCOPY
should still maintain the wavelengths of the copied subsection correctly.
Also it should be possible to use LISTPIXELS, IMPLOT, GRAPH, etc.
In order for this to work requires all tasks in IRAF to agree on a
representation of the coordinate system in the image headers and to use
a system interface for accessing coordinates.My initial work tried to preserve use of the APNUM keywords but it turned
out that the IRAF WCS interface requires use of the WAT (world coordinate
attribute) keywords for specialized applications. I could go into more
detail but I don't think that is what is of interest. There are certainly
important drawbacks, the header is less readable and not easily accessible
via IMFORT. There are advantages too such as interfacing to other packages
and a more compact header. There is also an advantage that the SPECn
attributes are extensible to store nonlinear dispersion functions which is
not possible with FITS cards which are limited to 80 characters. Note that
things are more complicated because other WCS keywords must be folded into
the computation of wavelengths to account for any subsampling, origin
changes, subsections, changes in dimensionality (a 1D section of a 2D
echelle image), etc.There are only two things I can tell you. The new software can understand
the APNUM keywords, i.e. the old format, as input for backwards compatiblity
with previous reductions. However, any new spectra created have the new
WCS structure. The other thing is for the case of a single linearized
spectrum in an image a simpler, readable, format is possible and you can
go between the two using SCOPY. I suspect that neither of these is that
useful for you.It is not going to be possible to reinstate the old format except as
noted above for input of the older format. What should be done is to
provide a WCS interface in something like IMFORT. Even then it might
be complex if you are doing more than just accessing wavelengths and
need to update things.Use of NEWIMRED is certainly possible but there may be some pain in
switching between NEWIMRED and other packages. If you would care to
specify what types of things involving access to the APNUM keywords
you do maybe I could provide suggestions, an IMFORT interface routine,
or a SPP conversion task. The main concern is what you need to do
in the way of updating the keywords and not simply getting
info about the aperture and wavelengths.Sorry for the surprise and difficulties. There are many important
reasons for the changes which many people will like but one casulty is
IMFORT access.Yours,
Frank