What matters is if they’re separate programs or not. They’re separate programs. Systemd is, like, 25 different pieces of software. It’s not 25 objects that are dynamically loaded or something. No, it’s 25 different pieces of software.

Most distributions only use some of them, and the “some” set is different from distro to distro. Many distros swap parts out. And some distros only use systemd-resolved.

Systemd is, legitimately, well divided. Most of the complaints against systemd are from people who know basically nothing about systemd.

Why does it matter that it's 25 binaries on disk? Surely the logical structure matters way more than the filesystem structure. Is Windows 3000 pieces of software because there are 3000 executables in system32?

It matters if they're separate programs or not, and they are. You can pick and choose, and distros do. Systemd is not a monolith, it just isn't.

Is Windows 3000 pieces of software?

How many pieces of software is nginx or Apache, where every feature is an optional plugin?

Are they working together with interdependencies or not?

E.g. explorer.exe is not replaceable on windows so the display server and that is one software. (The kernel is NOT though, because it can surely run in a headless way). It's not that hard of a concept..

the display server and explorer.exe on windows are much more separated than systemd-init and systemd-journald...

Well, does the way spreadsheet cells are handled has any relevance to presentations? If not, then they don't increase each other's complexity budget.

systemd-init has nothing to do with network resolution.

Except it does, because it signals systemd-resolved whenever a network interface service comes up or goes down. And where do those come from? They're created automatically by systemd-init's tight coupling with the network subsystem, which it shouldn't have.

systemd-journald is a nicer example though.