The OpenCPU System: Towards a Universal Interface for Scientific Computing through Separation of Concerns

Jeroen Ooms. Preprint on arXiv.
Conceptual, high level introduction to embedded scientific computing and the OpenCPU system. Highlights the domain specific problems of integrating statistical software in systems, pipelines and applications, and emphasizes the importance of the prinicple of separation of concerns.

Enforcing Security Policies in R Using Dynamic Sandboxing on Linux

Jeroen Ooms. Journal of Statistical Software, vol 55(7), 1-34, 2013.
On security and allocation of hardware resources in the context of computing services. Explains why standard approaches to security might not work for statistical software. Introduces a package to enforce security and hardware restrictions using mandatory access control and rlimit on Linux.

Directions for Improved Dependency Versioning in R and CRAN

Jeroen Ooms. The R Journal, Volume 5/1, June 2013.
On management of the decentralized open source development process in the context of statistical software and specifically R. Argues how proper dependency management is key to reliable applications, stable repositories and reproducible results.

RProtoBuf: Efficient Cross-Language Data Serialization in R

Dirk Eddelbuettel, Murray Stokely, Jeroen Ooms. Under review
Protocol Buffers are a method of serializing structured data between applications. They offer a unique combination of features, performance, and maturity that seems particularly well suited for data-driven applications and numerical computing.

A Practical and Consistent Mapping Between JSON Data and R Objects

Jeroen Ooms. Under Review.
A set of conventions is proposed for converting between R and JSON objects, and implemented in the R package jsonslite. The mapping defines how JSON data is parsed and generated from R objects in the OpenCPU API, but has many other applications as well.