Scientific instrumentation is the lifeblood of experimental science. However, novel instrumentation that is developed in research laboratories is often not reproducible. Not only are the detailed designs and assembly information rarely published with instrumentation papers, this knowledge is often not adequately record by the original research group. As staff members move on the knowledge of how to recreate this information is lost. The inability to recreate instruments and experimental setups from published work helps drive the reproducibility crisis in science. We present a new software and data structure for capturing assembly instructions and design information as an instrument is built, in a semi-automated and machine readable way. The data structure can also describe multiple configurations of the same instrument with minimal data duplication. This program was designed to aid the open sharing of scientific hardware designs as research data, but would equally be applicable for preserving institutional knowledge of custom instrumentation.

Preprint of journal article. Abstract: Collaborative design of physical products between remote partners poses unique challenges. This is due to both the complex and interconnected data required for product design and manufacture, and to the centralised computing infrastructure traditionally used to manage product lifecycle data. While modern cloud based solutions to collaborative design are gaining popularity, they diminish the control of each design partner. In contrast, software designers readily collaborate on highly complex software, while retaining direct control of the files they are editing due to the dominance of distributed version control. This version control can be coupled with “Developer Operations” or DevOps tools to automate critical processes and facilitate communication. In this paper we explore how DevOps workflows can be adapted to the development of hardware. We include concrete examples of how this can be implemented in practice from a case study of the OpenFlexure Microscope project. While much ground remains to be broken in this field, we believe DevOps for hardware can support a new paradigm of distributed hardware development, with enormous benefits for both commercial and open-source hardware.