search

pyspeckit / paper1

First paper about pyspeckit
https://pyspeckit.github.io/paper1/main_compressed.pdf
1 stars 7 forks source link

Comments #4

Open migueldvb opened 6 years ago

migueldvb commented 6 years ago 
Introduction

but \texttt{IRAF} development ceased xx years ago and is currently only
supported in Python 2.7 by
AstroConda\footnote{http://astroconda.readthedocs.io/en/latest/index.html}.

From a look at the IRAF github repository it seems that all the issues and pull requests are left unanswered. Other long-standing problems with IRAF have been the very complicated installation procedure, and many issues in the licensing which made attempts to package the code by distributions fail. Even though in the last release from 2012 they claim that all license restrictions have been removed, there is still some code from Numerical Recipes that is not freely distributable.

The command language for IRAF in Python (PyRAF) has support for python 3. I have been maintaining the python 2 and 3 versions of the package in archlinux for a while. I don't know why astroconda recommends python 2.7 apart from support for STSDAS which is an external package.

astropy integration

Unit conversions in \pyspeckit are now (as of October 2015) done entirely using
the \astropy system.

This sentence may sound out-of-date and could be replaced by.

Unit conversions in \pyspeckit are done entirely using the \astropy system
since \pyspeckit v0.1.16 released in May 2015.
keflavich commented 6 years ago

Thanks. This stuff would have been fine as a PR too. Could you give me some supporting information to back up the IRAF statements? i.e., who can I say is maintaining PyRAF?

See commit e90dffa

migueldvb commented 6 years ago

That looks great, thanks.

Here is a summary of the licensing situation at the time of IRAF v2.12.

http://iraf.net/forum/viewtopic.php?forum=4&showtopic=138501

There were several licenses, including some of them non-free. In particular, several NCAR routines in the graphics infrastructure were under a proprietary software license and a modified version of the yacc source code from the 1970's was under the UNIX source license:

    This directory contains the source for the Yacc compiler
    compiler (Stephen C.  Johnson), as modified to produce SPP
    language parsers.  You should have a UNIX source license to
    use this software on your machine.

A list of files in IRAF v2.16 that contain code from Numerical Recipes and patches using LAPACK and FFTPACK from the Debian maintainer are here:

https://github.com/iraf/iraf-v216/pull/37

Pyraf is maintained at the STScI and has now a stand-alone repository at https://github.com/spacetelescope/pyraf. Python 3 support was added around early 2013. I think it is more accurate to refer to the upstream projects in the paper rather than what astroconda is packaging.