Your film journal, without the film school
A film journal is a working notebook rather than a diary. You watch something, then write about what it was doing: what worked, what you noticed, what would have gone past you two years ago. Jonas Mekas ran a column under that name in the Village Voice from 1958 to 1975, collected by Columbia University Press. Film school assigns a smaller version of the same exercise. Editors and screenwriters keep private ones for decades, because naming what you saw is a different mental act from seeing it.
Outside that professional context almost nobody sustains one past the second month. This gets blamed on discipline. It is better explained as a scheduling error built into the format.
The quota problem
A film journal, as usually described, is a per-title obligation. One viewing, one entry. That structure quietly assumes you have something to say every time.
You do not. Nobody does. Out of five viewings, maybe one leaves an actual thought behind: a structural choice worth arguing with, a performance doing something strange, the sense that the whole thing was built around a single image. The other four were fine, or were not, and afterwards there is nothing in you beyond that.
A per-title format therefore produces four forced entries for every real one, and the forced entries are what kill it. Writing three sentences you do not mean is unpleasant in a specific way. It trains you to associate the notebook with owing something rather than with noticing something. Within a few weeks the choice at 11:40pm is between writing hollow lines or skipping, and skipping breaks the continuity the format depends on.
The habit that would actually have worked has the opposite shape. Write only when there is something, accept long gaps, do not pretend the gaps are failures. But that cannot function as a journal on its own, because a notebook with entries for one viewing in five is not a record of anything. There is no year in it. Fifteen paragraphs, floating, with no idea what surrounded them.
That is the bind. The writing has to be occasional to stay honest. The record has to be continuous to be worth keeping. A single habit cannot be both at once.
Two jobs sharing one name
Split them and the difficulty stops being spread evenly across every night.
The log is the fact that you watched a thing: title, date, sometimes a rating. No creative content, no quality bar, entirely mechanical. Its only value is coverage, so it has to be complete.
The notes are the thinking. This has value only when meant, which means it has to be allowed to be rare.
Most setups fail by running these as one habit on one cadence. The mechanical half gets a reverence it does not need, and the thinking half gets a quota it cannot honour. Reverse both. Automate the log so coverage never depends on how you feel at midnight, and let the notes be voluntary. Then the paragraphs you meant land inside a documented year, and nothing is owed on the nights when there is nothing.
What that looks like over one month
Concretely, take an ordinary month with nine viewings in it.
Under the per-title format, you write nine entries or you write two and then stop. Usually the latter, and the two you wrote are for the first week, because that is when the notebook was new.
Under the split arrangement, all nine appear in your daily entries without any action at the end of any night, each one sitting next to whatever else that day held. Twice during the month something actually lands, and you write four hundred words about it because you wanted to. At the end of the month the record contains nine viewings and two pieces of real writing.
The second version is strictly better and requires less from you. That is unusual enough to be worth noticing.
The automatic half
deariary covers the log side. Connect Trakt, which is itself fed by scrobblers for Plex, Jellyfin, Kodi and others, and what you watched arrives in that day’s entry with no closing step. The Trakt integration guide goes through what is collected, including ratings and any review text written there.
The point for this argument is narrow: coverage stops being a function of your evening mood. A month where nothing landed is still documented. Something you half slept through and forgot by Tuesday is still in the record, which turns out to matter later precisely because you would never have logged it on purpose.
Scaffolding, in other words. Not writing.
What it does not do
It does not produce criticism, and a generated line about something you watched is worth nothing as criticism. It records that the thing happened, in context. Naming what you saw stays the part with any value in it, and automation has no business there.
It has no access to what you noticed, either. The entry knows a title and an hour. It does not know the whole thing is shot at eye level, or that the cut you are still turning over is forty minutes in. Those go into Trakt, where they get picked up, or into a notebook that is nobody’s business.
And it is not an archive of your own writing. If the goal is a searchable body of criticism you produced, free-form writing tools are the right category and this is not it.
The version that lasts
Stop asking one notebook to be both things.
Let the record accumulate on its own so the year exists. Write when something is actually there, which will be rarely, and let those pieces sit inside a year that did not need them.
In five years what you have is not a film journal you kept. It is one you have, which was always the part that mattered.
