The new engineer asked
how a release gets to production.
You said
you would write it up.
You knew it well.
You had used it
every week for three years.
Ten minutes, you thought.
You wrote the first step.
Merge to main.
Then a build runs.
What starts it?
Then the image is tagged.
By what?
With what?
Then it goes to staging.
Who promotes it?
Is that the same pipeline,
or a second one?
An hour later
you had a page of questions
and you were reading YAML
you had never opened.
You had not forgotten it.
You had never known it.
You had known
how to use it.
In 2002
Leonid Rozenblit and Frank Keil
at Yale
published a set of studies
with a simple design.
People rated
how well they understood
everyday things.
A zipper.
A flush toilet.
A cylinder lock.
The ratings were high.
Then they were asked
to explain how each one worked,
step by step,
in writing.
Then they rated themselves again.
The ratings fell.
Writing out the mechanism
showed people
how much of it
they had never had.
The researchers called it
the illusion of explanatory depth.
It is strongest
for things with parts that cause each other.
Devices.
Processes.
Pipelines.
The reason is plain.
Using something every day
feels like understanding it.
You can see the parts.
You know what goes in
and what comes out.
The feeling of knowing
is built from familiarity,
and familiarity
is not a chain of causes.
So test the feeling
before you lean on it.
Before you estimate a change,
before you say it is simple,
before you tell someone
how it works,
write the chain.
Not what it does.
How.
This happens,
which causes this,
because of that.
Every link you cannot write
is a place where you were guessing.
That is not failure.
It is the map
of what to learn next,
found before production found it for you.
And when someone else
says they understand it,
do not ask if they do.
Ask them to walk you through it.
– Serguey Asael Shinder
