Frequently Asked Questions
If your question is not answered below, please create an issue on GitHub and ask it there.
Can I help?
Yes ! You can:
- Become a reviewer by filling your information and watch for new submission
- Submit paper for the work you've already replicated
- Spread the word about ReScience in your community (twitter, mailing lists, blogs, etc.)
- Star the project on GitHub
- Help improving this website by forking it and propose modifications (through PR)
- Opening an issue if you have any unanswered question.
- The Website could certainly use the help of a Jekyll expert and the eye of a designer (but keep in mind we want to keep a simple & clean website where content is the primary focus)...
- Print stickers from the ReScience material repository (and possibly send some to us...)
What's the difference between replication and reproduction?
There is no concensus yet on what exactly these two terms mean, so here is how we understand and use them.
Reproduction of a computational study means running the same computation on the same input data, and then checking if the results are the same, or at least "close enough" when it comes to numerical approximations. Reproduction can be considered as software testing at the level of a complete study.
Replication of a scientific study (computational or other) means repeating a published protocol, respecting its spirit and intentions but varying the technical details. For computational work, this would mean using different software, running a simulation from different initial conditions, etc. The idea is to change something that everyone believes shouldn't matter, and see if the scientific conclusions are affected or not.
Reproduction verifies that a computation was recorded with enough detail that it can be analyzed later or by someone else. Replication explores which details matter for reaching a specific scientific conclusion. A replication attempt is most useful if reproducibility has already been verified. Otherwise, if replication fails, leads to different conclusions, you cannot trace back the differences in the results to the underlying code and data.
In the ideal world, the reproducibility of computational studies would be checked automatically before publication. In the real world, most published studies do not even claim to be reproducible, because the code or the input data are not made public. Given this situation, a ReScience publication provides two important services: by replicating the original work, it provides an independent implementation of the original computational protocol, and by making this new implementation public, it is reproducible and thus a safer basis for future research to build on.
Are there any publication fees?
No. ReScience promotes open access and relies on the volunteer work of editors and reviewers. And the free hosting by GitHub.
Can I submit a paper using proprietary tools such as Matlab?
No. If we aim at replicability, we have to make sure the proposed implementation allows anyone to re-run and understand the code without any barriers. Proprietary software constitutes such a barrier because of their price and their closed nature. An implementation specifically targeting a free alternative such as Octave will be accepted though.
Is it mandatory to have an open-source license on the code?
Yes. The code accompanying ReScience articles will be archived at Zenodo along with the article and data, entirely under open licenses. This is part of the open access principle underlying ReScience. For the choice of a license for the code, we refer to the the Debian Free Software Guidelines.
Can I submit the replication of my own research?
No. Mistakes in the implementation of research questions and methods are often due to biases authors invariably have, consciously or not. One's biases will inevitably carry over to how one approaches a replication. Perhaps even more importantly, we aim at the cross-fertilization of research and trying to replicate the work of one's peers might pave the way for a future collaboration, or may give rise to new ideas as a result of the replication effort.
Can I suggest a replication?
Yes. If you want to suggest an article for a replication, just open a new issue in the call for replication repository and give the reference of the original article and possibly the reason you would like to see this article replicated (please refrain from suggesting your own work). Note that you're also encouraged to register as a reviewer such that you can review the replication you've been proposing if someone actually takes up the challenge.
If you're looking for some challenge, you can also look at the current list of suggestions.
If my submission is rejected, can I resubmit it?
This depends on the reason your submission was rejected. If the original research has already been replicated previously (in ReScience or elsewhere), it is unlikely we'll consider another replication, even if you think your implementation is better. However, if your submission has ben rejected because reviewers cannot make it run, you are encouraged to re-submit once the identified problems have been solved.
What if I fail to replicate a research result?
Some research may not be replicable. Before declaring a research result non-replicable, we require extra caution to be taken. In addition to scrutiny of your submission by reviewers and editors, we will contact the authors of the original research, and issue a challenge to the ReScience community to spot and report (using the issue tracker) errors in your implementation. If no errors are found, your work will be accepted and the original research will be declared non-replicable.
Is there a preferred implementation language?
We favor a Python implementation because this language offers a scientific programming stack that can handle research in many domains of science (see for example the Scipy 2015 and EuroScipy 2015 conferences). Python is also often reported by scientists to easier to learn and use than other languages. Furthermore, using Python might allow us to use the IPython notebook and the upcoming Jupyter environments.
However, other languages or tools are acceptable if they are mainstream in the respective domain of research.
I'm a student, can I submit?
Yes ! Students are strongly encouraged to submit their work. Although the ReScience publishing model is a bit different from other academic journals, it can give students a first experience at peer-reviewed scholarly publishing, including meeting standards of scientific rigor and addressing reviewers' comments. Publishing in ReScience is also a way to actively contribute to open science while adding to one's publication record.
How do I know if my replication is successful?
This depends mainly on the original paper introducing the research results to be replicated. Results might be presented in a qualitative or quantitative way and you'll have to clearly state why you think your implementation of a method or protocol is a faithful replication of the original results.
What kind of research can I replicate?
Any computational research in any domain of science as long as there is an editor from the Board who has the expertise to edit your submission. The editorial board is growing to increase the scientific domains being covered. If no editor is able to edit your submission, you can also propose a guest editor (who must be willing to work with our GitHub-based editorial processes).
Does ReScience issue DOIs (Digital Object Identifiers)?
ReScience itself does not, but every article published in ReScience receives a DOI from, and is indexed in Zenodo.
What is the preferred format for the accompanying article?
We use exclusively Markdown format, from which articles are converted to HTML, PDF, and other formats using pandoc.
We realize that this may impose some limitations on formatting possibilities. However, it is the tradeoff we chose to be able to offer a green open-access journal. All editors and reviewers are volunteers, and we want to minimize their workload.