The ordinary Tuesday you will want back
There is a specific kind of day people describe wanting back, and it is never the day you would expect.
Not the wedding, which is on film. Not the trip, which is in a folder. The one people describe is an ordinary Tuesday: a weekday in a flat they no longer live in, with a routine they no longer have, in the company of someone who was still around then. Nothing happened in it. That is part of what makes it unbearable to have lost.
The interesting question is why that particular Tuesday became valuable when ten thousand others did not, and the answer has a structure worth understanding, because it settles an argument about how to keep a diary.
Value is assigned afterwards, by events the day knows nothing about
A day does not contain its own significance. Significance gets conferred later, from outside, by what happens next. There are a few recognisable ways this occurs.
The day before a line. Something changes: a diagnosis, a message, a departure, an accident. Everything before it becomes the last of its kind, and the ordinary Tuesday two weeks earlier acquires a status it did not have at the time. It is now evidence of how things were, which nobody needed evidence of until things stopped being that way.
The day that holds an earlier version of something ongoing. A friendship before it cooled. A city you were new in. A child at four. The relationship still exists, but the version of it that existed that Tuesday does not, and there is no other way to reach that version.
The day that turns out to be a beginning. You met someone at a work thing and thought nothing of it. You read something on a train and it changed how you worked for a decade. At the time it was nothing; a year later it is the origin of a large part of your life, and you cannot remember any of the detail.
The day that establishes the baseline. Only interesting once the baseline has moved. What a normal Tuesday consisted of in a period you now think of as difficult, or easy, or the best years, and which you can no longer describe with any accuracy because it was just normal.
In every case the property that makes the day matter is not a property of the day. It is a relationship between that day and a future that had not happened yet.
Which makes selection impossible in principle
The standard advice about journaling assumes you can identify the days worth recording. Write when something happens. Capture the meaningful moments. Do not worry about the boring ones.
If significance is conferred retroactively, that advice is asking for a prediction you cannot make. Not one that is hard to make. One that requires information that does not exist at the time of the decision.
You would need to know that in three weeks the phone call will come. That the person you had lunch with becomes central. That this is the last year in this flat. Nobody knows those things, and the days that turn out to matter are indistinguishable at the time from the ones that do not, because at the time they are the same kind of day.
This is why selective journaling produces the record it produces. Not a record of your important days: a record of the days you believed were important, which is a different and much less interesting object. It systematically captures milestones, which are the events already preserved elsewhere, and systematically drops ordinary Tuesdays, which are the only copy.
Coverage is the only available strategy
When you cannot identify which items in a set will be needed, and the cost of keeping an item is low, you keep all of them. This is not a philosophy. It is what you do with backups, and it is the same reasoning.
The reason nobody applies it to their own life is that the cost per day has historically been fifteen minutes of writing, and fifteen minutes multiplied by every day for a decade is an enormous bill to pay for an unknown payoff. At that price, selection is rational even though it does not work.
Change the price and the calculation inverts. If the marginal cost of recording a day is approximately zero, coverage becomes the obvious choice, and the prediction problem disappears rather than being solved. You do not need to know which Tuesday matters. You need to not be deciding.
That is the entire argument for automatic capture, and it is an argument about economics rather than about diaries.
The bad ratio is the point
Here is the part that sounds like a weakness and is not.
Most entries will never be read. Out of a year, you might reread twenty days, and nineteen of those will be for a reason you cannot currently anticipate. The hit rate is terrible.
For anything with a per-item cost, a terrible hit rate is a real problem. For something that costs nothing per item, it is irrelevant. Nobody complains about the photographs they never look at again, because taking them cost nothing. The one that matters, ten years later, was worth the entire folder, and no version of that folder curated in advance would have contained it.
A diary that only exists on the days you judged worth it is a curated folder. A diary that exists on all of them is the folder.
What the product is actually for
deariary writes an entry each day out of what your connected services logged, without asking you anything. The point is not productivity, not self-improvement, not analytics on your own behaviour.
The point is the moment six months or six years from now when you open an ordinary Tuesday and the day comes back: the flat, the routine, the person, the thing that was going on that you had entirely stopped thinking about.
Everything else in the product exists to make that moment possible. Automation is there because a record requiring daily effort will not survive to be read. Prose rather than dashboards, because a chart does not return a day to you. Coverage of every day rather than the notable ones, because notability is decided later and by then it is too late to go back and record it.
What it cannot promise
It cannot make an entry moving. Most days will read as a flat account of a flat day, which is what they were.
It cannot recover what you felt. The record holds the facts of a Tuesday, and the feeling arrives, when it arrives, from you meeting the facts again years later. Sometimes nothing arrives.
And it cannot reach back. Whichever ordinary Tuesday you already lost is lost. This only applies going forward, which is the least satisfying sentence in this post and the most important one.
The whole thing in one line
You cannot know which day you will want back, so the only workable strategy is to keep all of them, which is only feasible if keeping them costs nothing.
That is what this is for.
