[jerf]

Step one is to "break the ice" and get them interested in computers.

Step two is to figure out if they're any good at that.

A good childhood should be spent breaking the ice on numerous possible interests and finding out what works with you and what doesn't. Everybody gets abundant opportunity today to figure out whether they like sports, of all kinds; we could do with more (good!) opportunities to figure out if you like programming, or many of the other interests and hobbies that few children ever get exposed to. Most won't be any good at programming. That's OK, because everybody should be trying lots of things, and you won't be terribly good at most of them. (Where "good" here includes some concept of enjoying it enough to want to do it on your own, along with raw talent.)