Awakening: Modern Challenges and Opportunities of Software Engineering Research Conference: 2026 IEEE/ACM 48th International Conference on Software Engineering: Future of Software Engineering ; April 12–18, 2026; Rio de Janeiro, Brazil 2026 IEEE/ACM 48th International Conference on Software Engineering: Future of Software Engineering (ICSE-FoSE), April 12–18, 2026, Rio de Janeiro, Brazil DOI: 10.1145/3793657.3793901 ISBN: 979-8-4007-2479-4/2026/04 CCS: Software and its engineering CCS: Social and professional topics Funding CCS: Computing methodologies
Diomidis Spinellis, Zoe Kotti · 2026
Read Awakening: Modern Challenges and Opportunities of Software Engineering Research Conference: 2026 IEEE/ACM 48th International Conference on Software Engineering: Future of Software Engineering ; April 12–18, 2026; Rio de Janeiro, Brazil 2026 IEEE/ACM 48th International Conference on Software Engineering: Future of Software Engineering (ICSE-FoSE), April 12–18, 2026, Rio de Janeiro, Brazil DOI: 10.1145/3793657.3793901 ISBN: 979-8-4007-2479-4/2026/04 CCS: Software and its engineering CCS: Social and professional topics Funding CCS: Computing methodologies by Diomidis Spinellis, Zoe Kotti (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-21403.txt.
Introduction
Spinellis and Kotti argue that software engineering research faces a structural squeeze. For decades the field used open tools and systems that were easy to study at small scale. Many key systems today are large, closed, and rooted in industry settings that academics cannot reach. Small tweaks to method will not fix this, they say. They urge industrial PhDs, lasting industry partnerships, bigger teams, moonshot projects, and changes in funding and review.
Catalog record
| CREATOR | Diomidis Spinellis, Zoe Kotti |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22915 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.21403
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Awakening: Modern Challenges and Opportunities of Software Engineering Research Conference: 2026 IEEE/ACM 48th International Conference on Software Engineering: Future of Software Engineering ; April 12–18, 2026; Rio de Janeiro, Brazil 2026 IEEE/ACM 48th International Conference on Software Engineering: Future of Software Engineering (ICSE-FoSE), April 12–18, 2026, Rio de Janeiro, Brazil DOI: 10.1145/3793657.3793901 ISBN: 979-8-4007-2479-4/2026/04 CCS: Software and its engineering CCS: Social and professional topics Funding CCS: Computing methodologies Diomidis Spinellis, Zoe Kotti Abstract . Software engineering research benefited for decades from openly available tools, accessible systems, and problems that could be studied at modest scale. Today, many of the most relevant software systems are large, proprietary, and embedded in industrial contexts that are difficult to access or replicate in academia. We review how the field reached this point, identify structural challenges facing contemporary research, and argue that incremental methodological refinement is insufficient. We discuss practical directions forward, including industrial PhDs, long-term industry-academia collaborations, larger research teams, moonshot projects, and changes to funding and evaluation practices. † † cc-license: by 1. Introduction In Awakenings ( Sacks 1973 ) , neurologist Oliver Sacks reports on his work at a New York hospital treating patients with post-encephalitic Parkinsonism who had spent decades in extreme rigidity, mutism, and institutional stasis. The book presents detailed case histories of patients given L-DOPA in the late 1960s, a treatment that at first restored movement, speech, and initiative with striking clarity. These sudden awakenings were short-lived: dosage escalation led to severe side effects, emotional instability, and the gradual return of rigidity or new pathological states, forcing patients and clinicians to confront the limits of pharmacological recovery. Awakenings thus records not a cure but a transient reopening of consciousness. Here we argue that software engineering research has passed though a similar temporary happy age of innocence where development processes, projects, tools, and methods allowed the advancement of knowledge in a simple, effective, open, and reproducible manner. A victim of its own success, software engineering research is now conducted in an environment where the complexity of processes and artifacts is beyond what can be easily developed or studied in an academic laboratory. Researchers often have to choose between relatively trivial studies of unrealistic toy problems and multidimensionally challenging studies in proprietary contexts. 2. A Brief History of Software Engineering Research The following paragraphs trace software engineering research from its inception to the present, illustrating how early productivity driven by openly available tools, low-hanging fruit, and open-source software gave way to Big Tech’s research advantages. Birth in a Crisis. Software engineering emerged in response to a crisis of chronic overruns, unreliability, and escalating maintenance costs. The 1968 NATO Software Engineering Conference ( Naur and Randell 1968 ) crystallized this diagnosis and established an engineering framing focused on disciplines, methods, and measurement. The community institutionalized through dedicated conferences, starting with the 1975 meeting on reliable software ( Boehm and Hoare 1975 ) (often called “ICSE-0” ( Bianculli et al. 2018 ) ), where 22 of 53 papers came from industry and four from government. At the time, the raw material needed for rigorous research—large codebases, defect reports, effort data, tools, and operational experience—lived inside companies and was largely proprietary. This made early software engineering research unusually dependent on industrial access and anecdote-driven generalization. Tool Building. A defining force in early software engineering research was building tools for software development tasks. AT&T Bell Laboratories developed and published foundational technologies including filters ( Meunier 1995 ) , file differencing and archives, pattern matching, macro processing, language translation ( Kernighan and Plauger 1976 ) , programmer’s workbenches ( Dolotta et al. 1978 ) , language development tools ( Johnson and Lesk 1978 ) , build automation ( Feldman 1979 ) , configuration management ( Rochkind 1975 ) , static analysis ( Johnson 1977 ) , and domain-specific languages ( Bentley 1986 ) . Importantly, a 1956 consent decree ( Lewis 1956 ) prohibited AT&T from commercializing non-telecommunications products, so Unix and its tools were liberally licensed royalty-free ( Salus 1994 , p. 60) with widely available source code. This allowed universities worldwide to apply, study, and enhance these tools. Low-Hanging Fruit. Pinning down known-unknowns allowed low-hanging fruit to be collected. One strand concerned the tighter management of the software process. For example, Brooks 1975 performed early empirical action research on large-scale software development, while Humphrey 1989 introduced software process maturity models and levels, process assessment and change, and several software process elements we now take for granted: the project plan, configuration management quality assurance, software inspections and testing, data gathering and analysis, automation and defect prevention. Other influential researchers contributed to the economics of software development and its maintenance ( Boehm 1981 ; Boehm 1983 ) and the understanding of software evolution ( Lehman 1996 ) . On the product side, major advancements concerned structured programming ( Dijkstra 1968 ) , axiomatic program correctness ( Hoare 1969 ) , information hiding ( Parnas 1972 ) , program families ( Parnas 1976 ) , and model checking ( Clarke et al. 2009 ) . Open Source Software. The emergence of open-source software ( Androutsellis-Theotokis et al. 2011 ; Fitzgerald 2006 ) was a boon for software engineering research. First, it provided an avenue for disseminating and evaluating new software tools, ranging from configuration management systems ( Tichy 1982 ; Chacon and Straub 2014 ) , to unit testing frameworks ( Beck and Gamma 1998 ) , to static analysis tools ( Ayewah et al. 2008 ) . Second, it provided product and process data of key software, such as the Linux kernel ( Schach et al. 2002 ) , Eclipse ( Ying et al. 2005 ) , the Apache web server ( Williams and Hollingsworth 2005 ) , and MySQL ( Han et al. 2019 ) , with repositories that could be easily and profitably analyzed at scale ( von Krogh and von Hippel 2006 ; Bird et al. 2009 ; Kalliamvakou et al. 2016 ) . Third, it advanced software reuse, an early key software engineering goal ( McIlroy et al. 1968 ; Caldiera and Basili 1991 ) , through tens of thousands freely available (easily and a zero or little cost) open source components ( Spinellis 2007 ; Haefliger et al. 2008 ) efficiently distributed and deployed through package management systems ( Spinellis 2012 ) . Big Tech. Over recent decades, network effects ( Katz and Shapiro 1994 ) enabled by internet technologies ( Gallaugher and Wang 2002 ) provided virtual exclusivity to Big Tech in impactful software engineering research through control of the modern experimental substrate: hyperscale systems, telemetry, and real-world interventions at scale. Research within platforms (cloud, app stores, browsers, IDEs, CI, repo hosts) can be evaluated and shipped faster and more effectively than university prototypes. Frontier areas (ML systems, LLM tooling, security, reliability) are hard to study without industrial workloads and infrastructure, while tools are often provided opaquely through SaaS offerings, hindering open study. Recent research shows sustained Big Tech research growth and deep university co-authorship ties that concentrate agenda-setting power ( Mendonça et al. 2025 ) . Network effects amplify this trend in two ways. First, collaboration networks: once a few firms become hubs for high-impact work, top researchers and