I guess it depends what we consider “pacman”. At some point in the past, the pacman developers decided to split most of the logic into libalpm. Does that really mean that libalpm isn’t part of pacman? I struggle with not calling libalmpn part of pacman because most of what we think of as pacman has been moved into libalpm. If you look at the current source code for the pacman binary, it actually does very little except io and calls to libalpm. If the
pacman binary is the only thing we are calling “pacman” then is
yay no longer a pacman wrapper since it also uses libalpm directly?
It is true that Manjaro has discussed the possibility of splitting the pacman package so the pacman binary can be excluded but that conversation has been going on for over a year and pamac still depends on
pacman and not libalpm. Even if they did that, the difference between being a front-end to pacman or libalpm isn’t a very interesting distinction since all the package management logic is in libalpm.
The statement “pamac is a front-end for libalpm, not pacman” is completely accurate. I just think it downplays that fact that libalpm is the core of pacman. Further,
yay and others also use libalpm. If you are writing an application, why would you not want to interface with a library instead of parsing the results from a cli tool?