Thanks Mike, for explanation, I understand now better (after browsing the install script as well) the way how the installation is working. However my colleagues from other institutions have faced several problems connected with particular setup of their network. First: some have administrator not allowing to change the current setup (which was done by someone else including older Iraf at some computers) and they do not want the root account hacking in the system - some machines do not have iraf installed at all.
Nevertheless, the people would like to install IRAF there.Second: we have here in my department a special setup of a network where the Linux OS is NFS mounted from central server allowing each client individual /var /etc and /data directory but the /usr nad / is NFS mounted (in fact the kernel itself is obtained by TFTP using boot rom strategies. So we have iraf 2.12 on /rsoft/iraf and to be able to run it from here I have had to make many hacking editing paths in various scripts etc ... but I was still able to make links in root and /usrSo what I would suggest for the future is the possibility of a little degraded IRAF installation (without iraf networking - as I understand it from
http://iraf.noao.edu/faq/FAQsec09.html#9002 it is the main reason why to have fixed path to /usr/include/iraf.h and root /iraf link) which could be installed anywhere and run with the single script containing mostly setting of the environment variables as every proper unix program is doing.I am sure the image pipes in /dev/ are not so critical as well as I have seen many people using iraf only for data analysis (e.g. splot) on xgterm and when they need 2D image they use ds9 and its load image menu. (In addition it is easier to convince the adminstrator to create two files in /dev that do not modify the run of the system and where the previous IRAF installation is not in conflict)I really do understand the reasons for current iraf installation as a perfect complex system, however, I think the way of its "modus operandi" has changed from its design times. Most students simply would like to use IRAF for reduction or analysis on either ther own laptops (without networking, tape allocation and even direct printer connection - mostly creating PS files for inclusion in other documents) or their institutional computers in the large university campuses where the administartors do no meet requests for special system hacking - even not knowing what the astronomers may want.
The vast mojority of them will never compile their own packages.So an easy environment setup and directory untaring would be more than enough for their needs.Consider, please, such a option (for linux PC platform at least)
Best regards,Petr[quote:3a20468273]
The install script MUST be run if either there is no iraf already on the machine OR the iraf root is something other than /iraf/iraf. The script not only puts links in the system directories, it also edits paths in the iraf files themselves.
So, if the path changes the install script must be re-run.
Cheers,
-Mike[/quote:3a20468273]