3 ms·
For what it's worth, the OP argues the opposite - that systemd is insufficiently generic. Much of its complexity is a matter of conflating mechanisms and polic
by zozbot234 5y ago
For what it's worth, the OP argues the opposite - that systemd is insufficiently generic. Much of its complexity is a matter of conflating mechanisms and policy.
This is par for the course for what's essentially prototype-quality software that's trying to address a number of non-trivial issues - but let's not pretend that real improvement is not possible here, or that systemd-as-it-currently-is does not stray from what's often called 'the UNIX philosophy' or 'the art of UNIX programming', of which separating mechanism from policy is a key component.
- pas 5y ago> separating mechanism from policy as a key component of unix philosophy Can you please elaborate on this? It sounds like a very desirable approach, but I don't really see it "in unix". (I can think of a few program suites that use processes and piping between them as the "protocol", eg. cups, the executable files in a directory that are then launched in a loop pattern is more common, but that seems to lack the policy part. Whereas OpenSSH for example lacks this.)