Home: https://etherpad.wikimedia.org/p/oscibar2018
 
Session 10
 
Title: Software Citation

Moderator:
    Sophia Doerner
 
Participants (15): 
	* Alexander.Struck@hu-berlin.de @astruck
	* Michael Rustler (@MichaelRustler, Kompetenzzentrum Wasser GmbH, michael.rustler@kompetenz-wasser.de)
	* Guido (Scherp)
	* Dominique.Hansen@hu-berlin.de
	* Simon Worthington, TIB
	* Christian Pietsch (Bielefeld University Library) https://orcid.org/0000-0001-8778-1273
	* Stephan Druskat (HU Berlin, Citation File Format, @stdruskat, http://orcid.org/0000-0003-4925-7248)
	* Hannes Wünsche (FOKUS, WZB, WBI - Hannes.Wuensche@fokus.fraunhofer.de)
	* Konrad Förstner (University of Würzburg, konrad.foerstner@uni-wuerzburg.de, @konradfoerstner)
	* Daniel Mietchen (EvoMRI)
 
Summary:
    
Software as a subset of data is hardly starting to be recognized as a valid research result. The variety of programming languages, library dependencies and platforms (where software and data are mixed) only add to the complexity of how to properly refer to, respectively give credit for, digital research tools like software. We currently lack a consistent procedure of referral to such tools, and furthermore one which can improve findability and possibly even reproducibility in general.
In this session, current developments in software citation are introduced. The perceived lack of consistent processes (let alone standards) will be discussed. Best practice guidelines may be suggested / developed.

  
Notes:
    
Why do we even speak about software citation?
	* research practice / integrity / replicability
	* I want to get credit for the work I do when I'm not writing papers.

Is software data? In which aspects do they differ with consequences for citations?

Does it make sense to cite software the same way as articles? 

METRICS that would help evaluating software:
	* depsy indicator http://depsy.org/
	* GitHub stars, forks, downloads
	* CRAN downloads (but only from RStudio mirror): https://cranlogs.r-pkg.org/
	* citations for your software package or article (about the software)

Which priniciples should be adapted (metadata fields)?
	* Automation of container technology (e.g. Singularity, Docker)
	* R has a method that creates BibTeX data for the language R or a function https://cran.r-project.org/doc/FAQ/R-FAQ.html#Citing-R, similarly node.js
	* author (what about contributor, testers, managers, ) ... 
	* title
	* version
	* PID (or repo URL and commit hash)
	* programming language?
	
How is software currently cited or referred to?
	* indirect mentions
	* cited DOI from Zenodo (getting a PID for every release or every commit?)
		* Coupling Github with Zenodo: https://guides.github.com/activities/citable-code/
	* Write an initial article for journals like JORS or JOSS but later get your software cited by its PID, but we need to get these journals indexed in PubMed, Web of Science, BASEsearch, ...
	
Existing metadata schemas
	* CodeMeta https://github.com/codemeta/codemeta more formalized, more general metadata, Zenodo integration in the pipeline, not specifically written for software citation
	* CFF Citation File Format https://citation-file-format.github.io/
	* CiteAs ??
	
Where would you want to see software citation info/metrics?
	* GitHub? cite me as + license file
	* Zenodo?
	* re3data?
	* Software Hertitage?
	* Altmetric (https://www.altmetric.com/) ?
	
Multiple views to cite software
	* author: publish some sort of release that he/she wants to be cited
	* scientist: cite a specific version (not a release) available in the software repository
	* others given in https://doi.org/10.7717/peerj-cs.86

Recommendations:
	* As a developer: Add CodeMeta and/or CFF in you project (you can create a CodeMeta file from a CFF file soon) (but e.g. in case of R needs to be converted to offical "citation" file)
	* implement it as a container to be language independent
	* 

Link Collection:
    
https://github.com/danielskatz/software-vs-data https://peerj.com/preprints/2630/
https://doi.org/10.7717/peerj-cs.86 software citation principles
https://www.force11.org/software-citation-principles
https://github.com/codemeta/codemeta
https://codemeta.github.io/
https://citation-file-format.github.io/
https://cite.research-software.org/
‘Nature Software Submission Guidelines’, 2018. https://doi.org/10.1038/d41586-018-02741-4 or https://s3-service-broker-live-19ea8b98-4d41-4cb4-be4c-d68f4963b7dd.s3.amazonaws.com/documents/GuidelinesCodePublication.pdf. 
Ongoing list of links on software citation https://www.zotero.org/groups/1838445/o-s/items/tag/software-cite
Software Management plan: https://www.software.ac.uk/ssisearch?search_api_fulltext_1=software+management+plan
Citing the language or a package https://cran.r-project.org/doc/FAQ/R-FAQ.html#Citing-R
Wikidata is working on data models for software
	https://www.wikidata.org/wiki/Category:WikiProject_Informatics
	https://www.wikidata.org/wiki/Wikidata:WikiProject_ShEx
https://www.software.ac.uk/blog/2016-09-30-oh-research-software-how-shalt-i-cite-thee