I somewhat agree with him. I would, however, jump right into making something you would find interesting as soon as you have a basic understanding of what make

I somewhat agree with him. I would, however, jump right into making something you would find interesting as soon as you have a basic understanding of what makes up a packet, the TCP/UDP protocols, and how sockets work, such as an IRC chat clone, a simple networked game, a reusable self auto-updating library, and so on. I would recommend creating something that you'll actually use or remain interested in to stay motivated. I find I do most of my learning while trying to research my way around a stumbling blocks, but of course it always helps to read/research things beforehand as well.