Open-notebook science

Based on Wikipedia: Open-notebook science

The Lab Notebook That Anyone Can Read

Imagine a scientist working late in a laboratory, meticulously recording every observation, every failed experiment, every surprising result in a notebook. Now imagine that notebook is not locked in a desk drawer or filed away in a cabinet. Instead, it's posted online in real time, for anyone in the world to read. That's open-notebook science.

It sounds radical. It is radical.

In most of science, secrecy is the default. Researchers guard their unpublished data jealously, terrified that a competitor might scoop their results and publish first. The peer-reviewed paper—polished, complete, stripped of all the messy failures that preceded it—remains the currency of scientific careers. What you don't see in those papers is the ninety percent of science that didn't work: the experiments that led nowhere, the hypotheses that turned out to be wrong, the equipment malfunctions, the contaminated samples, the promising leads that fizzled out.

Open-notebook science throws all of that into the light.

No Insider Information

The phrase "open-notebook science" was coined in 2006 by Jean-Claude Bradley, a chemistry professor at Drexel University in Philadelphia. He described it with a simple slogan: "no insider information." If you're doing open-notebook science properly, everything the researchers know is available to everyone else. There's no privileged access, no early peek for collaborators, no embargo period before data goes public.

Bradley's definition was precise. There must be a web address—a URL—pointing to a laboratory notebook that anyone can access and that search engines can find. It doesn't need to look like a traditional paper notebook with lined pages and hand-scrawled notes. But it must contain all the information that the researchers used to reach their conclusions. Every data point. Every observation. Every misstep.

This includes what scientists sometimes call "dark data"—the experiments that failed, the results that seemed insignificant, the work that never made it into a publication. In traditional science, dark data disappears. It sits in filing cabinets, on forgotten hard drives, in the memories of researchers who eventually move on to other projects. Open-notebook science rescues that data from obscurity.

Why Would Anyone Do This?

At first glance, opening your notebook to the world seems like scientific self-sabotage. You're handing your competitors your ideas before you've had a chance to publish them. You're exposing your failures and blind alleys for everyone to see. You're making yourself vulnerable.

But the advocates of open-notebook science argue that this vulnerability comes with substantial benefits.

Consider the problem of reproducibility. Science is supposed to work because experiments can be repeated. If you publish a finding, other scientists should be able to follow your methods and get the same result. In practice, this often doesn't happen. Studies in psychology, medicine, and other fields have found that a troubling percentage of published results cannot be reproduced. Part of the problem is that published papers leave out crucial details. The methods section might say "we incubated the sample overnight," but not mention that the lab was unusually cold that week, or that the reagent was from a particular supplier, or that the researcher stirred the solution exactly seventeen times.

An open notebook captures all of that.

There's another benefit too: provenance. When data is fully open from the moment it's generated, there's no question about where it came from or how it was processed. In one example, a chemistry paper used spectroscopic data from three different batches of a compound to characterize a reaction. Each spectrum in the paper linked back to the specific experiment where that batch was created—labeled with designations like EXP099, EXP203, and EXP206. Anyone reading the paper could trace every piece of evidence back to its source.

The Broader Movement

Open-notebook science didn't emerge in isolation. It's part of a larger transformation in how research is conducted and shared, often called the open science movement.

The most visible element of this movement is open access publishing. Traditional scientific journals charge readers—or their libraries—substantial fees to access articles. An open access journal makes its articles freely available to anyone. The logic is straightforward: much scientific research is funded by taxpayers, so why should taxpayers have to pay again to read the results?

Closely related is open data, the practice of making the datasets underlying scientific papers freely available for others to analyze. A researcher studying climate change might publish not just their conclusions but also the raw temperature measurements, precipitation records, and atmospheric data they used. Other scientists can then check their work, try different analytical approaches, or combine the data with other sources.

Citizen science takes openness in another direction, inviting non-professionals to participate in research. Birdwatchers submit observations to ornithological databases. Amateur astronomers scan images for previously undetected asteroids. People install air quality monitors on their balconies and contribute readings to pollution maps.

Open-notebook science goes further than any of these. It's not just about sharing results or data or inviting participation. It's about sharing the entire process, in real time, hiding nothing.

Learning from Software

The open science movement draws significant inspiration from a revolution that happened in software development starting in the 1990s.

Open-source software is software whose underlying code is freely available for anyone to view, modify, and redistribute. The Linux operating system, which now powers everything from smartphones to supercomputers, was built this way. So were the Apache web server, the Firefox browser, and countless other tools that form the backbone of the modern internet.

The success of open source shocked a lot of people. How could volunteers, working without pay, scattered across the globe, produce software that competed with—and often surpassed—products from billion-dollar corporations? Part of the answer was that openness created a kind of distributed quality control. With thousands of eyes examining the code, bugs were found and fixed quickly. Good ideas could come from anywhere. The traditional corporate model, where code was a closely guarded trade secret, turned out to have hidden costs: isolation, duplication of effort, slower innovation.

Advocates of open-notebook science hoped that similar dynamics would apply to research. If experiments were visible as they happened, other scientists could spot errors early. Promising approaches could be adopted immediately. The wasteful duplication that happens when multiple labs unknowingly pursue the same dead ends could be reduced.

The Pioneers

Open-notebook science has attracted a small but dedicated community of practitioners across various scientific disciplines.

In chemistry, beyond Bradley himself, Matthew Todd at University College London has run an open-source malaria project, attempting to develop new treatments for the parasitic disease that still kills hundreds of thousands of people each year, primarily in sub-Saharan Africa. By making all research public, the project hoped to accelerate drug development and attract contributions from researchers worldwide.

In biology, Rachel Harding, a postdoctoral researcher at the University of Toronto's Structural Genomics Consortium, maintained a detailed public lab notebook while working on Huntington's disease research. Huntington's is a devastating hereditary condition that causes progressive deterioration of nerve cells in the brain. There is currently no cure. Harding's open approach meant that other Huntington's researchers could immediately see her results and build on them.

Theoretical fields have also embraced the concept, though their notebooks look rather different from wet-lab sciences. Carl Boettiger, working on computational modeling in ecology and evolution, kept his theoretical work open. Dror Bar-Natan, a mathematician, shared his research publicly. Daniel Himmelstein led a drug repurposing study that used GitHub—a platform originally designed for collaborative software development—as a kind of open notebook.

Some educational settings have adopted the approach too. At the University of New Mexico, a junior physics lab course used open notebooks, allowing students to learn not just the techniques of experimental physics but also the practices of open science.

The Fear of Being Scooped

Despite its benefits, open-notebook science remains rare. Most researchers have never seriously considered it. When they do, several objections tend to arise.

The most visceral concern, especially in biology and medicine, is what scientists call "being scooped"—having someone else publish your results first. Academic careers are built on publications. Getting scooped can mean losing credit for years of work. In competitive fields, where multiple labs race to make the same discoveries, this fear shapes everything: what researchers share at conferences, what they tell colleagues over coffee, whether they collaborate or compete.

Whether scooping actually happens as often as scientists fear is debated. Some argue that the anxiety far exceeds the reality, that most research isn't actually that competitive, that the fear itself creates a culture of secrecy that harms science more than the occasional stolen idea ever could. Others counter that in certain hot fields—the race to develop a new cancer treatment, the competition to publish a prestigious result—the threat is very real indeed.

Open-notebook science doesn't eliminate this concern. If anything, it makes your work maximally visible to potential competitors. The counterargument is that openness creates a clear timestamp: if your notebook shows that you made a discovery on a particular date, that's evidence of priority even if someone else publishes first. But this argument has limits. In the current system, what matters is the publication, not the discovery.

The Patent Problem

The second major objection involves intellectual property. In most countries, publicly disclosing an invention starts a clock ticking on your ability to patent it. In some jurisdictions, public disclosure immediately destroys patentability. Publishing your experiments in an open notebook constitutes public disclosure.

This isn't a hypothetical concern. Universities and research institutions often expect to patent discoveries, licensing them to companies that turn them into products. This generates revenue that funds more research. Researchers themselves may have financial stakes in patents derived from their work. For this kind of research, open-notebook science is simply incompatible with the business model.

Some advocates argue that the patent system itself is the problem—that patents on basic scientific discoveries often impede rather than promote innovation, that freely available knowledge produces more public benefit than monopoly rights. These are legitimate debates, but they don't change the immediate reality. If you're doing research that your institution expects to patent, you probably can't do it openly.

Can You Still Publish?

A related worry concerns traditional academic publication. If your results are already publicly available in your notebook, will journals still publish them?

The answer, surprisingly, is usually yes. Most scientific journals accept papers based on work that has previously been presented at conferences or circulated as preprints—preliminary versions posted online before peer review. Publishers have generally indicated, at least informally, that open-notebook science falls into the same category. Material from open notebooks has been successfully published in high-impact journals.

But there are exceptions. Some journals have strict policies against prior publication in any form. Researchers considering open-notebook science need to check the policies of the journals where they hope to publish. This creates uncertainty that may deter adoption.

The Correlation Problem

There's a subtler objection that scientists rarely discuss outside their own communities. Independent verification is one of the foundations of science. When two different research groups, working separately, reach the same conclusion, that's much stronger evidence than a single result from a single lab.

But what if the groups aren't truly independent? What if researchers in one lab read the open notebook of another and are subtly influenced by what they see? Not intentionally copying, but unconsciously making similar choices in how they design experiments or analyze data. Their results might appear to confirm each other while actually reflecting the same hidden biases.

This is called co-evolution of data, and it's taken seriously enough that large experimental collaborations sometimes establish "firewall" rules prohibiting communication between teams working on the same question. Open-notebook science tears down exactly those firewalls.

The Deluge

The final objection is practical rather than principled. There is already far more scientific literature than any human can read. Hundreds of thousands of peer-reviewed papers are published every year. Keeping up with relevant developments in a single subfield is a full-time job. Adding millions of unfiltered notebook entries to this flood seems overwhelming.

Advocates of open-notebook science acknowledge this concern. Raw notebook data isn't meant to replace peer-reviewed papers. It's a different kind of resource, useful in different ways. You wouldn't read someone's notebook cover to cover. You'd search it for specific details: exactly what protocol did they use? What did the raw data look like before processing? What variations did they try before settling on the published method?

But this requires good search tools, good indexing, good curation. It requires being able to distinguish high-quality notebook entries from sloppy ones. These are real technical and institutional challenges that the open-notebook community has not fully solved.

Degrees of Openness

Recognizing that full openness isn't possible or desirable for everyone, the community has developed gradations. Notebooks can be marked with logos indicating their level of sharing.

A fully open notebook is marked as "all content" and "immediate"—everything is shared, as it happens. A partially open notebook might share only "selected content," leaving out particularly sensitive experiments. Or it might impose a delay, sharing material only after a certain embargo period has passed. These variants are sometimes called "pseudo" or "partial" open-notebook science.

The distinction matters because partial openness loses some of the key benefits. If you're only sharing your successful experiments, you're not really illuminating how science works. If you're sharing with a delay, you've lost the ability for others to catch errors in real time or to build immediately on your results. The value of openness is not linear—the last ten percent may be the most important.

Infrastructure

Several platforms have emerged to support open-notebook science, with varying degrees of openness and features.

The Open Notebook Science Network is a WordPress blog network where researchers can create and maintain individual or lab notebooks. As of 2018, it had about 126 active members—a modest number reflecting how niche the practice remains. OpenWetWare hosts many laboratory groups and allows selective sharing of research information. Protocolpedia focuses specifically on sharing laboratory protocols—the step-by-step instructions for conducting experiments. Sci-Mate lets users define access permissions, offering flexibility between fully open and fully closed.

Beyond specialized platforms, some researchers have adapted general-purpose tools. GitHub, designed for software version control, works well for tracking changes to data analysis code and documentation. Zenodo, run by the European Organization for Nuclear Research (known as CERN), provides a repository where researchers can deposit and share datasets, papers, and other materials with permanent identifiers.

Incentives and Challenges

Open-notebook science has received some institutional support. The Open Notebook Science Challenge, focused on measuring the solubility of compounds in various solvents, attracted sponsorship from Nature (the prestigious scientific journal), Sigma-Aldrich (a major chemical supplier), and other organizations. Winners received recognition and modest prizes.

But institutional incentives mostly still point the other way. Hiring committees evaluate candidates based on publications in high-impact journals. Grant agencies assess proposals based on track records of conventional publication. Tenure decisions count papers and citations. None of these metrics reward open-notebook practice.

Until the incentive structures change, open-notebook science will likely remain a minority practice—adopted by those with secure positions, strong convictions, or research programs where openness offers clear strategic advantages.

A Different Way of Working

Perhaps the most significant barrier to open-notebook science is psychological rather than practical. Traditional scientific training emphasizes polished presentation. Papers are carefully written and rewritten. Data is cleaned and curated before anyone sees it. Admitting uncertainty or error feels like professional weakness.

Open-notebook science requires a different mindset. You have to be comfortable showing your work in progress, including the parts that are messy or confused or wrong. You have to accept that strangers might judge your half-formed ideas. You have to trust that the benefits of openness outweigh the costs of vulnerability.

This is a significant cultural shift. Science, for all its rhetoric about objectivity and error correction, is practiced by humans who care about status and reputation. Asking scientists to expose their mistakes in public is asking a lot.

Yet some researchers find liberation in openness. No longer maintaining a facade of constant success, they can engage more honestly with colleagues and the public about how science actually works. Their notebooks become not just research tools but teaching resources, showing students the reality of laboratory life: the frustrations, the dead ends, the occasional thrilling breakthroughs.

The Question Remains

More than fifteen years after Jean-Claude Bradley coined the term, open-notebook science remains an experiment itself. The hypothesis is that radical transparency improves research—accelerating discovery, increasing reproducibility, reducing waste. The evidence is still being gathered.

For most scientists, the costs still seem too high and the benefits too uncertain. They'll keep their notebooks closed, their data private until publication, their failures hidden from view. Science will continue to work, more or less, as it has for decades.

But the open science movement keeps growing. Preprint servers have become mainstream. Open access journals have proliferated. Data sharing requirements have become common. The culture is shifting, slowly, toward transparency.

Open-notebook science represents the logical endpoint of that shift. It asks: if openness is good, why not be completely open? If sharing accelerates science, why wait? If we learn from failures, why hide them?

These are questions without obvious answers. But they're worth asking. The scientists who practice open-notebook science are, in their own way, running an experiment on how science itself should be done. The results won't be known for years, maybe decades. But the notebook is open, and anyone can watch.