"A depends on B": what exactly are we talking about?

In programming, when we say “A depends on B”, it can mean a huge number of different things.

What the compiler sees

• A cannot exist without B.
• A knows B’s name.
• A has to hand data to B.
• A receives data from B.

All of that is visible directly inside A’s code.

What only a written test will see

• A has to run after B.
• A writes to the same place as B.
• A depends on B’s state.

All of that shows up when A and B are used together.

What potentially nobody will see (or that the customer will discover?)

• A and B have to interpret a value the same way.
• A and B have to use the same constant.

This kind of thing tends not to show up at all. It is the most dangerous.

And the list is not exhaustive.

Being precise

It matters to be precise. Whether you are discussing with other developers, or prompting an AI (which, unfortunately, or fortunately, cannot read your mind).

When you talk about a dependency, always ask yourself:

• on which plane? (source, execution, data, deployment…)
• in which form? (name, type, structure, meaning, timing…)
• who sees it? (the compiler, a test, nobody)

It is the kind of good reflex to have, whether you are designing a system, coding by hand, prompting, or doing a review.

The “source” plane is the one that a program’s dependency layers carve up.

Happy designing!