nonCOM

nonCOM is an open‑source, cross‑platform component model that serves as an alternative to Microsoft’s Component Object Model. It was introduced in 2008 by a collaborative effort between the University of Strathclyde and the open‑source hardware foundation to provide a neutral, language‑agnostic interface system capable of running on Windows, Linux, macOS and embedded environments.

The architecture of nonCOM retains key COM concepts such as reference counting, interface identifiers (IIDs), and a marshaling protocol, but replaces the proprietary binary registration mechanism with a textual interface definition language (IDL). The nonCOM compiler processes the IDL to generate platform‑specific binding code for languages including C, C++, Python, Rust, and Java. This approach enables developers to describe services and dependencies in a portable form and automatically create bindings for multiple languages.

Adoption of nonCOM has spread across several open‑source and commercial projects. It is used within the LibreOffice productivity suite, integrated into the Qt application framework, and supports the ROS‑2 middleware stack in automotive applications. The project is maintained through a public Git repository, with releases documented on the official website and a series of tutorials aimed at both beginners and experienced developers.

Current releases include nonCOM 1.5, which introduced support for asynchronous interface calls and improved garbage‑collection heuristics. The project is distributed under the LGPL v3 license, and the community actively contributes new interface modules and performance optimisations.

Critically, while nonCOM offers greater portability and a simpler licensing model than COM, its runtime performance tuning options are limited compared to mature commercial component models, and its ecosystem of third‑party components remains smaller. Nevertheless, its ongoing development, alignment with the Open Component standards, and support for a wide range of languages and platforms make it a viable choice for modular software architectures.