[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [fedora-astronomy] Iraf fork?



Iraf has been developed by NOAO until 2006, when the user support was
transferred to a volunteer effort, iraf.net. That was about version
2.12. In 2007, NOAO recollected the patches and user contributions,
and together with other improvements, released version 2.14.

Iraf is vast and ancient.  It's partly written in C and partly written
in SPP, a custom made language. SPP needs a custom compiler that it's
included in the iraf core. SPP is converted into fortran (using f2c)
and compiled with the C compiler.

Iraf works in different architectures, but lacks x86_64 support.  A
project[1] exists to port iraf to x86_64, and it's basically the work
of one person, Chisato Yamauchi[2]. This work it's outside NOAO or
iraf.net. I help him doing the testing of his developments in my
Fedora system, as he develops in CentOS.

Iraf includes a copy of the libc of the time of its early developments
(mid-80s), replicating the functionality of basic system calls. It
also includes a modified copy of ncar[3] graphics library 1.0 (current
version is 5.0). ncar wasn't free at that time (I don't know if it's
free know, it seems to be composed of several packages, some free and
some not).

Iraf graphics facility, x11iraf, contains non-free code also. I tried
to package it in the past[4], but the package couldn't be included due
to legal restrictions again. That's where the thread[5] in iraf.net
comes from. It was me complaining about the license text in the
iraf.net faq no representing the real status of iraf and x11iraf,
misleading into believing that both were free software when they
weren't. You can see also the answer to the complains about closed
code in iraf, basically "please fix it yourself and send me the
patches". As iraf is very far from trivial to compile, I suppose that
the number of patches received across the years is nearly zero.

Neither iraf.net or noao provide tools to work in iraf development,
such as a version control system. So, if we want to develop something
based in iraf, we at least need to establish a source tree somewhere.
Don't call it a fork if you want, but it's the only way to work.


[1] http://sourceforge.net/projects/iraf64
[2] http://www.ir.isas.jaxa.jp/~cyamauch/
[3] http://ngwww.ucar.edu/index.html
[4] https://bugzilla.redhat.com/show_bug.cgi?id=249614
[5] http://iraf.net/phpBB2/viewtopic.php?t=86793


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]