Quoting Simon Walter (simon@???):

> What I don't like is tight coupling.

I'm with you on that. This is why I see the problem as being pretty
much the entire Freedesktop.org 'stack', e.g., upower, udisks2, D-Bus,
udev, and all the rest -- not just systemd. I see all of them as being
problems, because they're all managed as CADT-style projects (unstable,
forever being EOLed and rewritten from scratch, etc.), create problems
for other codebases, and collectively form a hideous dependency
hairball.

> What I wanted to know about was why *Debian* decided to use it as
> the primary init program.

Honestly, the project did not decide. It happened by unplanned
incrementalism driven by the GNOME systemd-logind madness.

> In the above post, the author, 'dasein', mentions GR. Does this mean
> General Resolution?

Yes.

> Further on in the thread 'dasein' says:
>
> "If we ignore the people who preferred the relatively neutral option
> 2, we see from your own tally that 148 people preferred 'coupling is
> fine', and 95 'coupling is unacceptable' - that seems to be about
> the most direct way of measuring the size of the two poles to me,
> though obviously it doesn't tell you whether they're voting on the
> principle of maintainer autonomy, or on systemd specifically."
>
> Does this mean that there was a vote? Do you call that a simple majority?

It was a complex alternative vote using the Condorcet voting algorithm,
a variety of ranked-choice voting aka 'instant runoff voting'. The
(IIRC) four choices on the voted ballot were, in turn, the result of
some tactical dogfighting whereby the original question got amendments.

https://en.wikipedia.org/wiki/Condorcet_method