A ship's log is not a diary. The distinction matters more than it sounds like it should. A diary records what someone felt like writing down. A log records what the watch requires: position, heading, speed, weather, any event that bears on the safety or conduct of the voyage — entered at fixed intervals, in a fixed format, by whoever holds the watch, whether or not anything happened worth remembering. The log does not care if the entry is interesting. It cares that it exists, that it is timestamped, and that the next watch can pick it up cold and know where the ship is and why.
This is a strange kind of writing to do well, because it fights the natural instinct of anyone holding a pen. The instinct is to write when something happens and stay silent when nothing does. A log inverts that: the silence is exactly the information you cannot afford to lose. If the 4 a.m. watch skips its entry because the sea was calm and nothing seemed worth recording, the next officer inherits a four-hour gap with no way to tell whether it was uneventful or unrecorded. Those look identical from the outside, and only one of them is safe to assume.
Naval logs, aviation logs, and lab notebooks converge on the same answer to a problem that has nothing to do with any of their specific domains: how do you make a record trustworthy to someone who was not there? Not persuasive — trustworthy, which is a lower and more useful bar. The answer, in every case, is to constrain the writer. Fixed fields. Fixed cadence. No retroactive edits, or if corrections are necessary, a visible strikethrough and initials rather than an erasure. The rule in most maritime traditions is that you never tear a page out of a logbook, even the pages you wish you hadn't written. An error entered in good faith and later found to be wrong is corrected in place, dated, and left legible under the correction. The wrong entry stays readable specifically so that the correction can be checked against it.
This produces something odd: a document that is more credible because it is allowed to contain mistakes. A polished narrative, written after the fact with the benefit of knowing how things turned out, is exactly the kind of document you should trust less, because it has been shaped by information the writer didn't have at the time. A logbook full of provisional entries, later corrections, and the occasional admission that the 2 a.m. reading looks wrong on review — that is a document shaped by not knowing, which is the actual condition under which the events occurred. The imperfection is evidence of honesty, not a flaw to be edited out.
A flight data recorder takes the logic further by removing the human from the writing entirely. It cannot decide an approach was routine and skip logging it. It cannot smooth over a moment of confusion in the cockpit into a cleaner story once the flight has landed safely. Every parameter is written whether or not anyone thinks it matters, because the entire value of the device depends on not trusting anyone — pilot, airline, regulator — to decide in the moment what will turn out to have mattered. Investigators reconstructing a crash routinely find that the critical seconds were not the ones anyone would have flagged as significant while they were happening. The near-miss five minutes earlier that everyone forgot because nothing came of it. The instrument reading that looked like noise and was actually the first sign of the real problem. You cannot know in advance which four seconds of a four-hour flight will matter. So you record all of them, and let the question of what mattered get answered later, by someone with more information than you had.
This is the part of the discipline that generalizes furthest beyond aviation: judgment about what is significant is least reliable exactly when you are inside the events you're judging. The lab notebook tradition in experimental science encodes the same insight differently — you record the failed run, the reagent that seemed fine but wasn't, the result that didn't fit the hypothesis, not because you expect to need it but because the alternative is trusting your future self's memory of which failures were the boring kind. They rarely are. The failure that gets waved off as "probably just contamination" is, often enough, the one that turns out three months later to be the actual finding.
None of this is free, and the traditions that do it well are explicit about the cost. A ship's log takes a watch officer's time every hour, every voyage, for years, to produce a record that is almost always never read closely by anyone. Most flights generate recorder data that is overwritten and never examined by a human being. Most lab notebook pages document experiments that worked exactly as boring as expected. The discipline is sustained not because most entries turn out to matter but because you cannot know in advance which ones will, and the cost of maintaining the habit is much lower than the cost of the one time it would have mattered and wasn't there.
This is a bet, and it is worth being honest that it is a bet: you are trading a small, certain, recurring cost against a large, uncertain, occasional one. Institutions that keep this bet running for decades — navies, aviation regulators, research labs with real reproducibility standards — have generally concluded it pays off, but they had to decide that once, deliberately, against the pull of every individual moment in which the entry felt unnecessary. The log is not natural. It has to be chosen, and re-chosen, against the current of "nothing happened, why write it down."
It is worth being precise about what this discipline is not. A log kept for the benefit of the person keeping it — so that a future watch, a future investigator, a future version of the same scientist can trust what happened — is a different object from a record kept to monitor someone from outside. The format can look identical: timestamped, fixed-field, unedited. The difference is who it serves and who chooses to keep it. A ship's log is written by the officer who will also need to read the previous watch's entries to do their own job well; the discipline exists because the writer has a stake in the record being good, not because someone else is checking up on them. That distinction — a record kept for the sake of continuity rather than a record kept for the sake of oversight — is easy to blur and worth keeping separate, in any domain where logs get kept.
The watch officer who writes an honest entry at 4 a.m. is not performing compliance. They are doing a favor for the next watch, who is, in every sense that matters, a stranger to them — someone who wasn't there, who has only the page to go on, and who will have to make real decisions based on what it says. Good logs are written for that stranger. That is most of what makes them good.