The purpose of this Editorial is to announce a new table of contents section for the Journal of Chemical Physics (JCP). Following the success of our recent Special Topic Issues on Electronic Structure Software and Classical Molecular Dynamics Simulations, we have decided to encourage the submission of further papers describing computer software that can be used to solve problems across the full breadth of chemical physics, ranging from strongly correlated electronic structure theory to biomolecular simulations. From this point on, these papers will be published in a new journal section on “Chemical Physics Software” and highlighted in a Collection on the journal website. The remainder of this Editorial summarizes the type of software papers we are looking for and how they will be dealt with by the Journal.
As discussed in our recent Editorials for the Special Topic Issues on Electronic Structure Software and Classical Molecular Dynamics Simulations,1,2 computational chemistry and physics software has had a huge impact on many fields of science and engineering. This success has influenced development of new methods in theoretical chemical physics: in many cases, the most successful theoretical methods are now those that have been implemented in widely available software, those that are straightforward for others to implement, and those that run efficiently on modern computer hardware. Complementing this success are some challenges. Many theoretical methods are not straightforward to implement in software due to their inherent complexity, numerical instability, or slow convergence. In addition, computational chemistry and physics programs can involve a huge amount of computer code (sometimes a million lines of code or more), and this introduces its own challenges: How can graduate students and postdocs get acclimated to such large software packages and make valuable contributions within their limited time in a research group?
Our goal is to welcome software papers that address these issues. We are interested in papers that implement previously unimplemented methods, papers that describe how to get around implementation challenges, and papers that describe improved algorithms that allow computations to run faster, or with fewer resources, or in a way that takes advantage of emerging computer hardware. JCP is also interested in papers demonstrating ways to simplify the development of computational chemistry and physics software and to make this development less time consuming and more sustainable (e.g., through interoperability or re-use of community libraries). Papers reporting innovative analysis tools or novel platforms enabling, e.g., automation and high-throughput simulation are also welcome.
Not all papers on computational chemistry and physics software are suitable for publication in JCP. Unsuitable papers include those describing straightforward implementations of well-known theoretical methods, those resembling portions of a software user’s manual, and those that simply announce new software and list its features. Papers should have a clear relevance to chemical physics. Features that would make a paper more suitable for JCP would include discussions of how the software grew out of new theoretical developments (with a careful review of the relevant literature), discussions of why a particular algorithm or approach was chosen instead of other seemingly promising alternatives, discussions of software design strategies, and perspectives on future developments.
Since the best chemical physics software packages are under constant development, we have decided not to ask for a “snapshot” of the code accompanying the paper to be deposited as the supplementary material. Instead, we shall require a link to a git repository, or a similar publicly accessible code repository, to be included in the paper in the case of open-source software,3 or a link to a website from which the software binaries can be downloaded if it is not open-source. This way, interested readers will have access to the latest version of the software accompanying each paper we publish in the “Chemical Physics Software” section, including the latest bug fixes. Finally, to ensure that the papers published in this section satisfy this requirement, and also all of the other requirements we have described above, we shall send them out to review with a customized letter to the referees, inviting reviewers to install the software and comment on its quality and ease of use as well as on the quality of the manuscript.4 We hope that this new section of the Journal will benefit the community by disseminating the latest information about chemical physics software design strategies, implementation techniques, and capabilities, and how these connect to recent theoretical advances.
REFERENCES
The version of the software that is stable at the time of publication can be indicated with a version tag, a release, or a snapshot associated with a DOI, e.g., a Zenodo record.
For commercial codes, we encourage authors to provide a demonstration copy of the code for evaluation by referees. The JCP staff will act as an intermediary to ensure the anonymity of the peer-review process in cases in which the software cannot be downloaded freely.
