Configuring a wireless network interface under Linux used to be something that was sufficiently unpleasant to drive new users away. In Linux there are often multiple ways to do something. Many times this is a good thing – when you are a Systems Administrator you want to have as many tools and possibilities at your disposal to do things.
Yet for some things, I argue that it is not an improvement. It can also be a sign of a lack of a comprehensive, elegant, well-planned solution.
As Linux becomes more and more developed, more utilities are developed to make it more user-friendly. However, I would also argue that progress in user-friendliness can be problematic when the new tools or utilities are developed and become integrated more deeply into systems which do not address the fundamental objectives in a comprehensive fashion, thereby adding unnecessary complexity.
One of the strong points of Linux has to me at least always been the fact that you could easily administer and configure the system from the command line. Contrary to what administrators without much experience might think, command-line administration of systems is in fact more straightforward and less complex than via graphical interfaces or complex systems which are often designed to make it “easier”.
When Network Manager first appeared it was very welcome as a means to make the configuration of network interfaces, especially wireless ones, easier. Yet behind the scenes of this tool lies an ugly mess.
The primary method for configuring wireless network interfaces relies upon a set of wireless tools created by a developer named Jean Tourrilhes. These include the command-line tool iwconfig
In order to use cnetworkmanager, you still have to kill any running instance of nm-applet. This seems ridiculous to me.
To actually use it, you run a command like:
cnetworkmanager --unprotected -C access_point
cnetworkmanager -C access_point --wpa-pass=password