PCMSolver: an Application Programming Interface for the Polarizable Continuum Model

The growing complexity of quantum chemical program packages requires that an appropriate strategy be devised to implement new features. Scalability is of paramount importance, but it has become clear that maintainability and extensibil-ity of the code play an equally important role in managing this growing complexity [1]. The use of a modular programming paradigm has been recognized as beneficial in many other scientific computing contexts. New features are isolated into libraries that can be accessed by host programs through a well-defined application programming interface (API). Computational tasks are thus implemented into separate, independent and interchangeable modules. These are developed and tested independently of the particular host program in which they will be used [2, 3]. A well-defined API clearly delimits the boundaries of the functionality offered, thus forcing a programming style and standardisation of the functionality, eventually. When coupled with open-source licensing, modularity is a proper step towards ensuring reproducibility of results from scientific simulation software [4]. Open-source development can fully leverage the benefits of widespread, cloud-based, free, code development services, such as hosted distributed version control systems (DVCSs), continuous integration, code coverage, static and dynamic code analyses, nightly regression testing, public issue tracking, code review and so forth: adoption of a modern code development workflow is easily within reach. It is, of course, true that the above mentioned services are not exclusive prerogatives of open-source projects. However, an open review process of scientific software can often help to establish reproducibility, extensibility and sustainability of the software ecosystem. Open-source software, modular development of new functionalities and full-fledged exploitation of DVCSs ensure a much larger scientific impact. Third-parties can easily contribute to the project: by improving the documentation, by reporting bugs or by actively extending the codebase with new functionality. The description of chemical phenomena in complex environments, such as solutions or proteins, is as challenging as it is important for an accurate prediction of a wide range of properties [5]. Quantum chemistry is nowadays a valuable tool: its