Research Software is Failing

The digital infrastructure that powers modern science is dangerously fragile. Critical research tools operate without sustainable funding, relying on volunteers and short-term grants. This isn’t just inconvenient—it’s a systemic threat to scientific progress.

The Scale of the Crisis

🔬 Dependency Crisis

Essential tools across every research domain—from EEGLAB in neuroscience to NumPy in data science—operate without guaranteed funding.

⚠️ Critical Fragility

The loss of just 1-2 key maintainers could collapse entire research ecosystems that millions depend on.

💔 Maintainer Burnout

Volunteer maintainers burn out from unglamorous but essential work that receives no institutional recognition.

Why Traditional Funding Fails

🎯 The Innovation Bias

Funding agencies reward shiny new projects, not boring maintenance. But keeping software running is just as critical as building it.

Grant Cycle Mismatch

Research grants last 2-5 years. Software maintenance is forever. When grants end, maintenance stops—and software breaks.

🏢 Career Disincentives

Universities reward faculty for publications, not software upkeep. Research software engineers have no clear career path for maintenance work.

Real Consequences

🚨 Research Disruption

When critical tools fail, entire research programs grind to a halt while researchers scramble for alternatives.

💸 Wasted Resources

Researchers spend months troubleshooting broken software instead of doing science.

🔄 Reproducibility Crisis

Broken or abandoned software makes reproducing scientific results impossible.

🚧 Barriers to Entry

Poor documentation and unstable tools exclude newcomers and underrepresented groups.

The Network Effect Problem

Research software doesn’t exist in isolation. Tools depend on dozens of other packages. When one component fails, it can cascade through the entire ecosystem. A single point of failure can break thousands of research workflows.

Example: The NumPy Near-Miss

NumPy—fundamental to scientific computing in Python—operated for years without dedicated funding despite supporting millions of researchers. Only heroic volunteer efforts and recent foundation support have stabilized it. NumPy is the exception. Most research software lacks such visibility and support.

Why This Threatens Science Itself

Science builds on previous work. When research software fails, it doesn’t just inconvenience current researchers—it breaks the chain of knowledge transfer that enables scientific progress.

The sustainability crisis in research software is a threat to the continuity of scientific knowledge.


There Is a Solution

The math is simple: just 0.1% of grant funding could solve this crisis completely. We know how to fix this—we just need to act.