[Python-Dev] Official citation for Python

[Jacqueline Kazil]

RE: Why cite Python….

I would say that in this paper —
http://conference.scipy.org/proceedings/scipy2015/pdfs/jacqueline_kazil.pdf,
where we introduced a new library, we should have cited Python, because the
library was based in Python. We were riding on the coattails of Python and
if Python did not exist, then this library would not exist.

(taking this a level higher)
Just as someone doing research (a specific application) should cite the
Mesa library. Without the good and bad that is Mesa, their research would
have taken a different form.

Since my Ph.D is on Mesa, I will be citing Python there.

I think for more insight we can look at who has cited some of Guido’s stuff…
For example:
https://scholar.google.com/scholar?cites=900267235435084077&as_sdt=20005&sciodt=0,9&hl=en

Does that help?
RE: Just like R - Versions

@Stephen
Are you suggesting major versions or minor versions?
RE: Guido’s prio works

Some of those have weight already. Should we be picking one those and
pointing people to that?
Final decision

I am going to the NumFocus summit for maintainers of Science Python
libraries next week. I believe that the Science Python community is where
the main audience for this is… correct me if you think this is a wrong
assumption.

I thought I could take two to three concrete formats and user test there
and report on how community members who would be using the citation feel.

Good idea? Bad idea?

On Sun, Sep 16, 2018 at 4:35 AM Stephen J. Turnbull <
turnbull.stephen.fw at u.tsukuba.ac.jp> wrote:

> Jacqueline Kazil writes:
>
>  > *As a user, I am writing an academic paper and I need to cite Python. *
>
> I don't understand the meaning of "need" and "Python".  To understand
> your code, one likely needs the Language Reference and surely the
> Library Reference, and probably documentation of the APIs and
> semantics of various third party code.
>
> To just give credit to the Python project for the suite of tools
> you've used, a citation like the R Project's should do (I think this
> has appeared more than once, I copy it from José María Mateos's
> parallel post):
>
>  > To cite R in publications use:
>
>  >   R Core Team (2018). R: A language and environment for statistical
>  >   computing. R Foundation for Statistical Computing, Vienna, Austria.
>  >   URL https://www.R-project.org/.
>
> I guess for Python that would be something like
>
> """
> Python Core Developers [2018].  Python: A general purpose language for
> computing, with batteries included.  Python Software Foundation,
> Beaverton, OR.  https://www.python.org/.
> """
>
> I like R's citation() builtin.
>
> One caveat: I get the impression that the R Project is far more
> centralized than Python is, that there are not huge independent
> projects like SciPy and NumPy and Twisted and so on, nor independent
> implementations of the core language like PyPy and Jython.  So I
> suspect that for most serious scientific computing you would need to
> cite one or more third-pary projects as well, and perhaps an
> implementation such as PyPy or Jython.
>
> Jacqueline again:
>
>  > Let's throw reproducibility out the window for now (<--- something
>  > I never thought I would say), because that should be captured in
>  > the code, not in the citations.
>  >
>  > So, if we don't need the specific version of Python, then maybe
>  > creating one citation is all we need.
>
> Do you realize that `3 / 2` means different computations depending on
> the version of Python?  And that `"a string"` produces different
> objects with different duck-types depending on the version?
>
> As far as handling versions, this would do, I think:
>
> f"""
> Python Core Developers [{release_year}].  Python: A general purpose
> language for computing, with batteries included, version
> {version_number}.  Python Software Foundation, Beaverton, OR.
> Project URL: https://www.python.org/.
> """
>


-- 
Jacqueline Kazil | @jackiekazil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-dev/attachments/20180916/915b6720/attachment.html>