Security by default, in software, means that the default configuration settings are the most secure settings possible, which are not necessarily the most user friendly settings. In many cases, security and user friendliness are evaluated based on both risk analysis and usability tests. This leads to the discussion of what the most secure settings actually are. As a result, the precise meaning of "secure by default" remains undefined.
In a network operating system, this typically means first and foremost that there are no listening INET(6) domain sockets after installation. That is, no open network ports. This can be checked on the local machine with a tool such as netstat, and remotely with a port scanner such as nmap. As a general rule, a secure network, is only as secure as the least secure node in the entire network.
If a program uses secure configuration settings by default, the user will be better protected. However, not all users will care about security and may be obstructed by secure settings. A common example is whether or not blank passwords are allowed for login. Not everyone can, or is willing to, type or memorize a password.
Another way to secure a program or system is through abstraction, where the user is presented an interface in which the user cannot (or is discouraged to) cause accidental data loss. This however, can lead to less functionality or reduced flexibility. Having user control preferences does not typically cause this, but at the cost of having a larger part of the user interface for configuration controls.