[SciPy-user] [OT] advice about publishing a scientific software

Robert Kern robert.kern@gmail....
Mon Jul 16 10:51:30 CDT 2007


massimo sandal wrote:

> - Minor one: Where to publish it? I am used to think about SourceForge
> as THE place for publishing open source software, however a number of
> people told me about Google Code. Basically what I want, if possible, is
> no-nonsense SVN access and some integrated mailing list/forum. Google
> Code seems good at that -it also gives automatically Wiki space, which
> is a major plus for me. What do you think about? Drawbacks? Licence
> issues (i.e. something like "since you publish it on Google Code the
> code is also owned by Google" tiny-letters clause?) Is there some other
> alternative?

If you want "no-nonsense," use Google Code. Sourceforge is a lot of nonsense. I
get incredibly frustrated trying to use the Sourceforge website or any of its
spawn on either side (downloader or developer). I don't believe there are any IP
issues; you own your code, not Google.

> Moreover, I'd like to know your opinions about the degree of
> "polishness" and other requirements for initial publication of a
> scientific software.

Meh. Release early and often. Having the code public is a good way to motivate
polish. Your code may crash or miss features, but it's code that wasn't
available before. You are making a positive contribution regardless.

> As for documentation, I am working on actual, user
> manual and hacking (mainly but not only focused on plugin writing)
> manual with tutorials etc.. The sw still lacks an installer but
> "installation" is braindead (provided that you have all the correct
> dependencies): put all the files in the same directory and run "python
> software.py" from where: do you think an installer is absolutely
> necessary or it is OK to let it go this way, and add the installer later?

For a Python project, you really should be using distutils, even if it is just
the simplest possible setup.py.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


More information about the SciPy-user mailing list