Why are there 2 different ways to ...
 

Why are there 2 different ways to have multiple IP addresses on a single interface? One of the ways is to simply add IP addresses to a single interface. The other is to create an alias interface name. These are clearly different structures inside the kernel to effectively do the same thing.

The method of using an alias interface name (e.g. eth0:1) is the most awkward to work with (because you have to search to find which interface name has which IP address to delete one, or find an available name to add one). So why can't that one be depricated and just removed in the next major Linux kernel? Both the "ifconfig" and "ip" commands support the simpler method.

Or am I missing some other advantage of interface alias names?

Ifconfig doesn't show the second address set by ip addr add .....

Oh, I guess it just does so for IPv6. They should fix that.