"A Lot of Moving Parts": A Case Study of Open-Source Hardware Design Collaboration in the Thingiverse Community

TL;DR

The paper investigates how open-source hardware projects collaborate within a collective design platform by presenting a detailed case study of DrawBot on Thingiverse. Using Thingiverse's public API, the authors assemble a longitudinal, networked dataset of 74 related Things and 567 relevant conversations, analyzed via qualitative coding to distinguish design-related and non-design-related contributions. They reveal a spectrum of collaborative activities (feature development, bug fixes, documentation, issue management, community support) and identify significant platform limitations, such as lack of proper version control and fragmented documentation, which drive workarounds like remixing and manual description updates. The study offers best practices for maintainers, design implications for platforms, and methodological guidance for OSH researchers, accompanied by a publicly available dataset to spur further CSCW-oriented OSH work.

Abstract

Open-source is a decentralized and collaborative method of development that encourages open contribution from an extensive and undefined network of individuals. Although commonly associated with software development (OSS), the open-source model extends to hardware development, forming the basis of open-source hardware development (OSH). Compared to OSS, OSH is relatively nascent, lacking adequate tooling support from existing platforms and best practices for efficient collaboration. Taking a necessary step towards improving OSH collaboration, we conduct a detailed case study of DrawBot, a successful OSH project that remarkably fostered a long-term collaboration on Thingiverse - a platform not explicitly intended for complex collaborative design. Through analyzing comment threads and design changes over the course of the project, we found how collaboration occurred, the challenges faced, and how the DrawBot community managed to overcome these obstacles. Beyond offering a detailed account of collaboration practices and challenges, our work contributes best practices, design implications, and practical implications for OSH project maintainers, platform builders, and researchers, respectively. With these insights and our publicly available dataset, we aim to foster more effective and efficient collaborative design in OSH projects.

Figures (6)

• Figure 1: Cumulative number of Things published on Thingiverse from January 2009 to November 2023.
• Figure 2: Overview of research methods, from data collection to analysis.
• Figure 3: Drawing Robot "Thing" studied in this paper. The metadata includes: (A) Thing name, author's username, and publication date; (B) Thing description, number of files, makes, and remixes; (C) user activity (e.g., likes, comments, collects); (D) the comment panel, with commenter usernames, posted dates, and threaded replies; (E) the remix panel, including remix activity with page links and author usernames.
• Figure 4: Directed network of Things within the DrawBot project. The direction of the arrows indicates a remix relationship, pointing from ancestor Thing to child Thing. The orange node represents the DrawBot Thing; yellow nodes represent DrawBot remixes; green nodes represent DrawBot ancestors; and blue nodes represent other Things within the network indirectly connected to the DrawBot Thing (i.e., 2 or more nodes away).
• Figure 5: Contributors to the DrawBot project. Each node represents a maker who made a design-related contribution, with the size of the node representing the number of contributions made (note: for the sake of figure legibility, we scaled down the central node since this was the DrawBot author, who made numerous contributions). The edges represent conversations (i.e., threads) shared between makers, with the edge weights showing the frequency of collaboration (i.e., the number of conversations). Blue, green, and red nodes represent exclusively hardware, electronics, and software contributions, respectively. Multi-coloured nodes represent interdisciplinary contributors.