
When a Lab Notebook Becomes Patent Evidence — and When It Doesn't
Two lab notebooks sit side by side on the same exhibit table. Both were kept by diligent researchers, in the same organization, across roughly the same stretch of a development project. The court leans on one of them. The other gets skimmed and set aside. In Korea, lab notebook patent evidence splits into these two fates long before any judge reads a line of chemistry — and what divides them is not the quality of the science. It is whether the notebook can answer the only two questions a courtroom ever asks of a record: when was this written, and has anything changed since?
The uncomfortable part is how little of a notebook’s evidentiary strength has to do with how good the research was. A brilliant experiment recorded in a way that cannot be dated is, evidentially, close to an experiment that never happened. A plodding series of failures, recorded so that each day proves itself, can outweigh it easily.
Where a Notebook Actually Enters a Patent Dispute
The most common misconception is that a notebook is how you win a patent. It isn’t. Korea operates a first-to-file system: the patent goes to whoever walks through the patent office door first, not whoever had the idea first. A notebook cannot win that race, and researchers who assume it can are often the ones most blindsided later. What the notebook does is decide what happens afterthe race is over — in at least three distinct situations.
The prior-user defense
Korean law recognizes a prior user’s right (Patent Act Article 103): a party that was already practicing an invention in Korea, or clearly preparing to, before someone else’s filing date — and that arrived at the invention without knowledge of the other party’s work — may keep operating under a non-exclusive license. Notice what the right does not do: your earlier invention date does not defeat their patent. At best it buys you a narrow permission to keep running your own process. And even that narrow permission must be proven: what you had, and when you had it. The record that has to carry this burden is almost always the research notebook.
Who really invented it
Employee-invention compensation disputes that climb all the way to the Patent Court rarely turn on whether the invention was valuable. They turn on who conceived what, in which order, and how much each person contributed. Development notebooks get submitted in these fights as a matter of course. The ones with provable dates actually move the needle. The ones without get argued around — by both sides, in opposite directions, which tells you what the court ultimately does with them.
What the company held, and when
Trade-secret claims need to establish what information existed, at what point in time, and under what degree of secrecy. A notebook is the closest thing a company has to a ledger of its own knowledge. If the ledger cannot be dated, the company is asking the court to reconstruct its memory from the memories of interested witnesses.
Three different situations, one shared demand: the notebook has to function as a clock. Which is precisely where things fall apart.
Why Can’t the Notebook Prove When It Was Written?
Korean courts do not reject research notebooks outright. Under the free evaluation of evidence, a notebook is in principle admissible. The real battle is over weight. Admissibility asks whether the document can be considered at all; weight asks how far it should move the court’s conviction. A notebook with undated entries, or dates added in one batch at the end of a project, or passages “reconstructed” after the dispute began, fails the second question, not the first. The document is allowed into the room and then quietly ignored.
The most damaging pattern is also the most innocent-looking: backfilling. Notes written in a single sitting — one pen, one hand, uniform ink, covering three months of work. They look meticulous. But their uniformity is the problem: nothing about the record itself ties any line to the day it describes. Three months of research, provable only as “sometime before the lawsuit.”
Digital files make this worse, not better. A file’s creation and modification dates are set by the same machine, and editable by the same hand, as the file itself. To a court weighing tampering risk, a document created in March 2024 and a document backdated to March 2024 are the same document. Timestamps that live inside the thing being authenticated are not timestamps; they are formatting.
The Revision That Erases Itself
Paper had a strange, underappreciated virtue: it kept its wounds. A struck-out value, a correction in the margin, a page re-copied with the original still bound in — every honest correction carried its own visible history. You could see what changed, and get a rough sense of when. This is why long-standing research-notebook guidance (including the administrative rules for national R&D projects) insists on objective facts recorded without falsification, participant by participant: the record is expected to show its own edits, not to hide them.
Digital records inverted this. Editing a file leaves no scar tissue. A silently overwritten value, a deleted row, a replaced figure — the current state gives no testimony about any previous state. And once tampering cannot be excluded, the record’s testimony about “what we knew at the time” collapses, because “at the time” is exactly what the revision trail existed to preserve. A notebook without an intact history of its own changes is not strong evidence with a flaw. It is a statement, unsigned and undated, waiting to be characterized by whoever’s expert goes first.
Why Did Researchers Stamp and Sign Every Page?
Older corporate R&D labs ran on a ritual: every notebook page dated, signed by the writer, then countersigned by a colleague who had read it — domestic guidelines, including the patent office’s own research-notebook handbook, recommended that confirmation signature come within a week or two of the entry. To the researchers performing it, the ritual felt like bureaucracy. It was actually doing one precise, load-bearing job: a third party was attesting, in ink, that this page existed in this form on this date. A distributed, human timestamp.
The ritual was largely inherited from the American first-to-invent world, where — until the America Invents Act took effect on March 16, 2013 — the question “who invented it first?” could decide ownership outright, and ownership battles were won and lost on witnessed, dated pages. Korea’s own system never needed that mechanism; it has been first-to-file since its modern beginning. But the practice crossed the Pacific anyway, embedded itself in corporate IP guidelines, and took root as orthodoxy. It was, for a first-to-file country, an oddly imported habit — and a genuinely useful one.
Then it died, quietly. Binders gave way to word processors, word processors to shared drives, shared drives to whatever came next — and the date-sign-countersign step simply did not survive the change of medium. Nobody ever decided to stop proving when things were written. The tool that proved it just stopped existing, and the habit’s ghost lives on in compliance manuals whose instructions no longer name a physical act anyone performs.
The Gap Nobody Inherited
Look at what the stamp-and-signature ritual was actually carrying. Not approval of the science. Not even accuracy. Just attestation of existence at a point in time — this page, this form, this date, seen by someone other than its author. Every modern workflow is drowning in dates: file systems, chat logs, document histories. But almost none of that was designed to be evidence, and it shows the moment someone with an incentive and a file system editor gets involved.
The stakes are not hypothetical. In national R&D projects, records of this kind already decide who gets sanctioned when things go wrong. In patent disputes they decide narrower but harder things: whether you may keep running a process you thought was yours, who the real inventor was, and what the company actually owned on the day it claims to have owned it.
The question stopped being “did you write it down?” Everyone writes things down now, more than ever. The question became whether the record can prove when it was written and that nothing has changed since — and most records in existence today cannot. That is the gap the old ritual left behind. For the moment, nothing is standing where it used to.