← Back to Home

How to Read man Pages

2026.07.29 · Linux

When I first started reading about Linux commands, man pages felt long and messy — I never knew where to start. Eventually I learned a reading flow, and it turned out to work not only for ordinary commands but for driver development docs too.

Common Symbols (SYNOPSIS)

Whenever you open a command's man page, the SYNOPSIS section tells you the basic syntax. It uses a few common symbols:

SymbolMeaningExample
[]Optional[-v]
<>Placeholder — replace it with a real value<file>, <pid>
|Either/or
...Can be repeated<file>...
--Start of a long option, or stop of option parsing--help, --

The Reading Flow

Every time you open a man page, don't try to read it all at once. Go in this order:

① NAME        → one line telling you what the command does
② SYNOPSIS    → learn the basic syntax: what's required, what's optional
③ DESCRIPTION → understand how it works and how it behaves
④ OPTIONS     → look up the specific options you need

Example

Take git blame. Its SYNOPSIS looks like this:

git blame [options] <file>

So the full usage is:

git blame main.c          # see who changed each line of main.c
git blame -L 10,20 main.c # only look at lines 10–20
📌 Scope
This reading approach isn't only for git, tar, or ls.
The same method also works for the man pages you'll use constantly in Linux driver development — ioctl, epoll, pthread, systemd, and more.
You can read them all with this one approach.