Re: Citation format for Octave

[rik]

On 05/26/2015 02:24 AM, Juan Pablo Carbajal wrote:
> On Sun, May 24, 2015 at 4:53 AM, Carnë Draug <address@hidden> wrote:
>> On 12 May 2015 at 00:03, rik <address@hidden> wrote:
>>> John, Carnë, and anyone else who regularly writes academic papers,
>>>
>>> What should the citation() function return for Octave 4.0?  The current
>>> text of the file is:
>>>
>>> -- CITATION --
>>> To cite GNU Octave in publications use:
>>>
>>>   John W. Eaton, David Bateman, Søren Hauberg, Rik Wehbring (2014).
>>>   GNU Octave version 3.8.1 manual: a high-level interactive language for
>>>   numerical computations.
>>>   CreateSpace Independent Publishing Platform.  ISBN 1441413006,
>>>   URL http://www.gnu.org/software/octave/doc/interpreter/
>>>
>>> A BibTeX entry for LaTeX users is:
>>>
>>>   @book{,
>>>     author    = {John W. Eaton, David Bateman, S\oren Hauberg, and Rik
>>> Wehbring},
>>>     title     = {{GNU Octave} version 3.8.1 manual: a high-level
>>> interactive language for numerical computations},
>>>     publisher = {CreateSpace Independent Publishing Platform},
>>>     year      = {2014},
>>>     note      = {{ISBN} 1441413006},
>>>     url       = {http://www.gnu.org/software/octave/doc/interpreter},
>>>   }
>>>
>>> We have invested a lot of time and effort in creating GNU Octave, please
>>> cite it
>>> when using it.  See also `citation pkgname' for citing Octave packages.
>>>
>>> -- End CITATION --
>>>
>>> This is clearly not right, but just changing the version to 4.0.0 isn't
>>> going to make the ISBN number correct.  Should we drop the publisher and
>>> note fields entirely?
>>>
>> We have an entry on the FAQ about this (which should be updated now).
>>
>>     http://wiki.octave.org/FAQ#How_can_I_cite_Octave.3F
>>
>> We ask users to cite the software itself.  Since not every journal will allow
>> this, the recommendation is to cite the published book if not possible.  But
>> citing software is becoming more common, and there will be fewer issues if
>> users try to cite it nowadays.
>>
>> There's this article [1] from the Software Sustainability Institute that
>> discusses how to best cite software (this is aimed at the users not at
>> developers, but it's a good view on the subject).
>>
>> Roughly, recommendation is to cite the software.  It specially recommends
>> against citing a paper or version of a manual (like R does), because it's
>> not version specific.  The following are some points from section 7 of the
>> article:
>>
>>     * In the first draft of a paper, always put software citations in
>>       references or bibliographies.
>>     * Be prepared to debate with reviewers why you have cited the software:
>>       you want to acknowledge the contribution of the software's authors and
>>       the value of software as a legitimate research output.
>>     * If a reviewer disagrees with a formal software citation, you can still
>>       make a general reference to the software in the paper.
>>     * If the software has a DOI (digital object identifier) use it to cite
>>       the software. If the software has its own website, use the website's
>>       URL for the citation.
>>
>> I think we should recommend.  Have a ready to use reference for the
>> software itself and a DOI (we can have one for each Octave release).
>>
>> And I guess we could some more public service and have citation() tell users
>> at the end that at very least, they should note the version of software used
>> in text, not for the sake of Octave developers, but on the interest of
>> reproducible research.
>>
>> Carnë
>>
>> [1] http://software.ac.uk/so-exactly-what-software-did-you-use
> Indeed, as pointed before citing software, so far, hasn't been a
> problem for me, here an example
>
> http://journal.frontiersin.org/article/10.3389/fncom.2013.00191/full#B69
>
> It would be nice to have a DOI. This was indeed was asked by the
> editors of that journal, but it was not a reason for removing the
> citation.

It would be nice to have a DOI, but I think we will just have to settle for
a URL.  According to the Wikipedia entry for DOI, "Organizations that meet
the contractual obligations of the DOI system and are willing to pay to
become a member of the system can assign DOIs."  I don't think we want to
use development funds just to join and get a single number.

--Rik

>
>