iraf.net

Hi Vassilis (and Bill),> > Question for you....when I start up xiraf, ds9 comes up with
> > stdimage=imt1024, and I cannot display
> > a full 2048x2048 image with the display command in IRAF, even if I
> > reset stdimage=imt2048. However,
> > I can display the same image correctly if I load it directly using
> > the File, Open button in DS9 itself.
> > Can you tell me how to fix this?
>
> I guess you've fixed that. Also, check that you've got a /usr/local/lib/imtoolrc symlink
(pointing to iraf$dev/imtoolrc) on your system. If missing then DS9
may be falling back to some default since it doesn't know then what an
'imt2048' is. You can also copy dev$imtoolrc to ".imtoolrc" in your unix
login directory if you can't write to system dirs.
> > (I also noticed a similar problem: when using the display command,
> > the coordinates of objects in the
> > image are not shown in DS9. But again if I load the image directly
> > in DS9, the coordinates are given.)
>
> This is an inherent limitation of IRAF in the way it interphases with
> the image display as it uses the fifos to pass the info on each
> pixels (X,Y, and pixel value). However the astrometry information is
> stored in the header of the fits file and this is not sent by IRAF to
> DS9 (or SAOimage or XImtool). I suspect that this happens for
> historic reasons. In the past IRAF could not display fits files, and
> each images was split on the header (.imh) and binary (.pix) part.
> May be then it was programatically difficult to send the WCS info
> with the "display" command. Actually, this is a limitation in DS9. Since the V2.11.3 days
(about 6 years ago) the display protocol will pass in additional wcs
info that includes a fully qualified path to the file. In XImtool this
info is used by an external task (which knows how to read the image/WCS)
to map display cursor positions to real-pixel/WCS values in the display.
Given the filename DS9 could do something similar to read the WCS from
the image and provide the same capability. I know they actually get this
information since they used the IIS display protocol code written for
ximtool verbatim, they just don't use it. It wouldn't matter if we sent
the entire header, DS9 would still have to be changed to use it! SAO has
been aware of this for a while but just hasn't fixed it.
Cheers,
-Mike