How "ps" Works And Why

Legacy Misunderstandings and the Evolution of ps

When you first start a Unix or Linux system, one of the first utilities you encounter is ps, the process status command. On the surface it looks simple: type ps and you get a list of processes. But in practice, the command is a minefield of quirks that trip up even seasoned sysadmins. Each Unix variant - Solaris, AIX, HP‑UX, Linux, FreeBSD - has its own flavor, and the flags, column headers, and even the way options are parsed can differ. This variability has long frustrated users who rely on scripts that run across multiple platforms. A classic example is the infamous ps -aux versus ps -e -f debate. On some systems, ps -aux is interpreted as a BSD style command that lists all processes and shows the full command line. On others, the same string is parsed as a Unix‑style option set, producing a different output or an error. The lack of consistency makes it hard to write portable code or even to debug a single command on a remote machine.

Many people, myself included, have been misled by the way ps behaves on their own platform. Back in 2001, I wrote a blog post titled