iraf.net

Don,
It sounds to me like you have, in principle, the necessary hardware
and software components to make a port possible. It would indeed be a lot
of work. Our experience has been that out of many IRAF ports attempted
outside NOAO/Steward, only one came close to success. Many others have
spent a lot of time working on the port only to give up when they encountered
serious runtime bugs that required machine-level debugging and intimate
knowledge of IRAF internals to diagnose and fix. An additional difficulty
in your case is that the various software components are less well integrated
than usual. For example, I would expect that magtape i/o to that ODI
controller would require custom work, and you would be in unknown territory
with the operating system, the C compiler, and the Fortran compiler [before
even thinking about beginning, make sure the C and Fortran procedure calling
standards are the same, so that you can call Fortran from C and vice versa].
Note that IRAF is much more than the usual application program or collection
of programs. It is a complete software development environment with its own
compiler/preprocessors, virtual operating system with memory and file
management etc. and thus the port involves some low level things one might
not normally encounter.

As for whether you could build just parts of the system: you could
build any subset of the core system or application *tasks*, but in order
to build even the set you mentioned you would have to build the entire
set of system facilities: the host system interface (HSI), the virtual
operating system (VOS), and some set of utilities like graphics kernels.
In terms of the work required to do the port, you still have to do the whole
system.> ... What I'm asking is,
> 1) is the project do-able at all;
> 2) if it is, can I do the whole system, or should I try only a part of IRAF;
> 3) if I decide to give it a try, where do I start?1) Although I can't say for sure, it is probably technically doable by an
experienced UNIX systems programmer who is familiar with machine level
debugging, but it would not be easy and usually we tell people they would
be spending more time on a risky port attempt than when converted into
money would allow them to purchase an entry level Sun.2) Have to port the whole system except possibly at the last minute, not
necessarily all of the executables, which wouldn't cut much work off the
port itself.3) You write back and ask for me to edit one of our existing port cover
letters and request from Jeannette a port tape, which you have to pay for
(unless you have access to the internet, in which case you can get the
port archive for free; I assume from your bitnet address you don't have
internet access). Please be sure you want to do this. I have put a lot
of work into those port letters in the past, and every bit of it has
turned out to be wasted!Cheers, Steve (rooke@noao.edu)

Don,
If you have IRAF on your VAX, and if it is running a sufficiently recent
version of VMS (v5.? later than 5.0 I think; I can find out for sure if you
need me to), and if you have ethernet between the VAX and your PC, you should
be able to use the PC like an Xterminal to get image display and vector
graphics out on the PC. VMS/IRAF v2.9x was distributed with saoimage and
xterm executables for vms. If you have ethernet, and your X server on the
PC can be used from X clients running on the VAX, what you would do is bring
up X on the PC, then use some terminal emulator to telnet to the VAX (use
xterm if you have it). On the vax, set display for the pc node (and anything
else in the readme instructions for saoimage), fire up an saoimage in the
background, then fire up the CL, reset stdimage=imtool, and you can display
images. If you don't have the necessary network hardware and software, it might
be cheaper to purchase it than to attempt an iraf port.Steve