They have an example service on the website:

(define sshd
  (service
    '(sshd ssh-daemon)                ;the secure shell daemon
    #:start (make-inetd-constructor   ;start on demand
             '("/usr/sbin/sshd" "-D" "-i")
             (list (endpoint
                    (make-socket-address AF_INET INADDR_ANY 22))
                   (endpoint
                    (make-socket-address AF_INET6 IN6ADDR_ANY 22)))
             #:max-connections 10)
    #:stop (make-inetd-destructor)
    #:respawn? #t))

(register-services (list sshd))
(start-in-the-background '(sshd))

Let’s see how the same service looks like with systemd:

[Unit]
Description=OpenSSH Daemon
Wants=sshdgenkeys.service
After=sshdgenkeys.service
After=network.target

[Service]
Type=notify-reload
ExecStart=/usr/bin/sshd -D
KillMode=process
Restart=always

[Install]
WantedBy=multi-user.target

I have some lisp knowledge, so the scheme version doesn’t look frightening to me, but I guess for sysadmins, who should write these kind of files frequently systemd’s TOML like language is much more easier to understand.

Some differences I see: Shepherd does some firewall management with ports, and I don’t see the services it depends on.

Why this kind of files should be written in a programming language at all? I guess it’s a remnant from the old times, but I like when tools abstract away the programming parts, and users shouldn’t have to deal with that. I like the same thing in docker-compose: I can configure a program whatever language it’s written, I don’t have to deal with what’s happening under the hood.

I guess there is some usefulness with defining services as code, if you need more complex situations, but it should the more rare case nowadays.