← Notes

Advice from outside is a hypothesis, not a diagnosis

This is an opinion about method: I think a written remark does not decide architecture, and that what decides it is data from the system itself.

I will put this as an opinion, because that is what it is: I think a written remark, on its own, should not move architecture.

Not because whoever makes it is wrong. Most of the time they are right in general and without knowing whether they are right here. Someone suggesting a technical path sees a symptom and returns the remedy the literature associates with that symptom. It is useful, it is generous, and it is not a diagnosis — because a diagnosis requires looking at the patient, and the patient is my system, with my numbers.

The risk of treating a suggestion as a diagnosis is not doing something stupid; it is doing competent work in the wrong place. I build the pipeline properly, it looks good, and the user still waits the same because the time was being spent at the other end. Nobody makes a visible mistake along that path, and the only loss is the month.

What I argue for is a cheap step between the advice and the build: measure the part the advice points at. If the measurement confirms it, the advice becomes a plan with priority. If it contradicts it, I have gained knowledge of my own system, which was what was missing from the start.

That is why I say it does not depend only on what is written. It depends on empirical experience or on observation from data — and both of those are produced by me, at home, not by whoever walked past the door and noticed.