To my knowledge, most of the problems people have with systemd are because it replaced an extremely old and well known and understood system generally implemented with shell scripts and replaced it with a more complex and engineered project to get speed and extra features (pre-opening pipes and re-spawning listeners on the other end when a client want something from them, etc). In an effort to provide a more seamless boot, traditionally external projects ended up being subsumed (DHCP, and DNS resolution to support robust network support on boot).
I'm not sure if it was ultimately the right decision or not, but since to my knowledge almost every major distribution has adopted it, I assume it's viewed as worthwhile by qiute a lot of people within the distribution planning community at least.
I'm not really sure how any of that relates to this project specifically, so perhaps you're referring to some other aspect of it?
However, it's obviously preferable to want/use what's being shipped on the device to all recipients.
In this libre-oriented space in particular, what users are likely to want as a priority is the ability to easily reproduce the bits shipped on the device from source, and be able to restore the device to a state no different than shipped from the factory using those self-reproduced bits. The freedom to run some hacked up half-baked alternative stuff they're technically free to install should they wish, is not the top priority.
A major component of the value conferred by this freedom is that those shipped, reproducible bits, contain a desirable foundation for the community to converge on and iterate from.
Telling people "you can install whatever you want" before even delivering is signaling "potentially unusable, controversial vaporware is shipping, community fragmentation ahead" to anyone paying attention.
I'm not particularly averse to systemd/GNOME, but I do fear that it's probably too immature an ecosystem for production mobile use and it's unnecessarily threatening the project's overall success.
It would have been more prudent to collaborate with the Jolla/Sailfish folks and ship basically a secure Nokia N9 successor but with an easily reproduced and flashed image, which the N9 lacked, while the GNOME community grinds away on getting their stack mobile-ready for a potential future device. I don't know if the licensing could have been hashed out to get the N9 stack fully available in reproducible form, but I'm inclined to assume that if there were components legally obstructed from such distribution, it probably was less work to reimplement them than build all the components needed for a GNOME phone.
