Skip to content

Latest commit

 

History

History
209 lines (159 loc) · 10.1 KB

papers.md

File metadata and controls

209 lines (159 loc) · 10.1 KB

Process for Authoring Jupyter Related Academic Papers

This document describes the processes used to author Jupyter related academic papers, which are typically submitted to peer-reviewed journals. This does not include talks and posters submitted to conferences unless the conference also publishes written proceedings.

On one hand, the primary focus of Project Jupyter is on producing open-source software. We are very much focused on writing code that delights our users. On the other hand, Jupyter (and IPython) emerged out of an acadmic setting, where research and peer-reviwed publications are a primary focus. To this day, many of our core contributors have academic positions and active research programs. This emphasis is also reflected in the usage of Jupyter across all areas of scientific research. Thus, in spite of a focus on writing software for users, Jupyter remains a tool that is by and for researchers.

Given this focus on research, the ongoing academic careers of many Jupyter contributors, and our desire to have an impact on computing research, it is important for us to author and publish peer-reviewed papers on Jupyter itself. Authoring these papers in the context of a large, open-source community presents a number of differences compared to many research collaborations.

Principles

We have tried to create a process for authoring papers with the following general principles in mind:

  • Inclusivity and generosity in authorship. Hundreds of individuals have contributed to the different Jupyter subprojects. These contributions span code, design, documentation, discussions, giving talks/tutorials, etc. We want to be generous in extending authorship privileges and responsibilities to as many of these contributors as possible.
  • Clear, explicit criteria for authorship. While being generous, we want have concrete, specific and auditable criteria for including individuals as authors.
  • Openness. The process of authoring papers should be open in the same way as the rest of the project's work. Thus, all of our papers are written in the public on GitHub.
  • Accountability. Being an author on a paper is a privilege, but also involves responsibilities. The concrete processes described below elucidate these responsibilities.

We expect to author and submit papers to different types of journals. Different journals will have slightly different concrete processes that appropriately embody these principles.

Author ordering

In some academic fields, author ordering is used to implicitly communicate the level and significance of the individual's contributions. We strongly feel this is misleading and misguided. Because of this, all Jupyter papers will use the following author ordering policy:

  • The first author will be listed as "Project Jupyter";
  • Individual authors will follow in alphabetical order; and
  • An explicit statement about the ordering will be included in the paper.

Listing the first author as "Project Jupyter" is important as it means that in abbreviated citations, the author list will be "Project Jupyter, et al." rather than artificially showing the first alphabetical author's name. This is common practice in academic field such as high energy physics where there are large numbers of authors on papers.

Process for Journal of Open Source Software (JOSS)

The Journal of Open Source Software (JOSS) is a peer-reviewed, developer-friendly journal for research related software. JOSS is unique in that:

  • Each paper is associated with a single GitHub repository;
  • The primary research artifact in the software itself; and
  • The paper itself is a relatively brief markdown file in the software's GitHub repository, deferring details of the software to its own documentation.

It is our intention to publish a JOSS paper for each of our main, user-focused subprojects (Notebook, JupyterLab, JupyterHub, nbconvert, ipywidgets, etc.). All such paper should use the process described herein.

JOSS authorship criteria

Our authorship criteria for JOSS papers are derived from those of the ICMJE:

  • Substantial contributions to the conception, design, or implementation of software; this includes coding, visual design, documentation, testing, discussions and other such contributions; AND
  • Final approval of the version to be published; AND
  • Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Anyone who satisfies and is willing to accept these responsibilities and commitments is welcome to be an author on any of our JOSS publications.

JOSS process

The following process should be used to author any Jupyter related JOSS paper:

1. Someone agrees to be the Coordinator for a paper

A single JOSS paper will be associated with a single Jupyter repository. The Coordinator will usually be a senior project contributor, Steering Council member, or leader of that subproject.

The role of the Coordinator is to implement this process for that paper. They do not have to write the JOSS paper themselves, but they will work with the community to organize the writing of the paper.

2. Open an issue and announce the writing of the paper

The Coordinator should open an issue to organize the writing of the paper on that repository and then announce to the Jupyter Google Group that the paper is being worked on and that anyone can contribute (with a link to the issue).

3. Draft the paper in the repository

At this point, the Coordinator should organize the writing of the actual paper.

The JOSS Author Guidelines detail the format and requirements of a JOSS paper. In general this will involve creating a paper subdirectory in the root of the repository with two files: paper.md and paper.bib. The paper should be written using our standard processes for pull requests and code review.

3. Email potential authors

Once a final draft of the paper has been merged into the repository, the Coordinator should email all potential authors of the paper, inviting them to participate. Usually this will include at least two emails:

  • Individually email all contributors listed in the Git log of the repository. This list can be generated using the Git command git log --all --format='%cN <%cE>' | sort -u.
  • Email the main Jupyter Google Group to include folks not listed in the Git log.

In addition, the Coordinator should send personal emails to individuals, such as students, who are new to writing academic publications, explaining why we are writing a paper, and encouraging them to participate. These personal emails should cc the individual's mentor or advisor if applicable.

Both of these emails should include:

  • A verbatim copy of the above JOSS authorship criteria;
  • A description the tasks each individual needs complete to be an author on the paper;
  • A concrete deadline for completing those tasks (at least 2 weeks out); and
  • A link to the paper's markdown file in the repository.

The tasks each individual needs to complete are as follows:

  • Create an account on ORCID.
  • Submit a PR to the repository adding their name, ORCID id, and affiliation to the paper;
  • Provide any edits to the content of the paper;
  • In the comment of the PR, briefly describe their contributions to the work; and
  • In the comment of the PR, provide verification that they have read the paper and agree to its publication.

As the deadline approaches, the Coordinator should remind potential authors to complete their tasks.

4. Final submission

Once the deadline for authors to complete their tasks has passed, the Coordinator should make sure authors are listed in alphabetical order. The Coordinator can then submit the paper.

Process for traditional academic papers

Papers submitted to other (non-JOSS) journals, will usually be longer and take a more significant amount of time to author. They will also usually be authored in dedicated repository (under the jupyter-resources GitHub organization). Because of these factors, the process is slightly different.

Authorship criteria

The authorship criteria for non-JOSS papers is slightly different than for JOSS in that all authors are expected to actively participate in the writing and editing of the paper.

Our authorship criteria for papers are derived from those of the ICMJE:

  • Substantial contributions to the conception, design, or implementation of software; this includes coding, visual design, documentation, testing, discussions and other such contributions; AND
  • Drafting the work or revising it critically for important intellectual content; AND
  • Final approval of the version to be published; AND
  • Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Anyone who satisfies and is willing to accept these responsibilities and commitments is welcome to be an author on any of our publications.

The second of these responsibilities implies that all authors will actively participate in the writing of the manuscript. We recognize that not all co-authors will contribute equally and in the same way to the writing of a paper. Furthermore, for papers with large numbers of co-authors, we expect the main writing will be done by a smaller group of co-authors, with other co-authors contributing to editing work and discussions. However, in all cases active participation in some manner is still required.

Process

Before creating the detailed process for authoring non-JOSS papers, we want to try out the above process for JOSS papers and see how it should be modified for longer form papers. In general, we expect to closely follow the process used in the SymPy project to author this paper.