I figured the answer was something along those lines - the books will help but they can't save me from problems in my implementation.

I have plenty of time for this so I'm going to persue the select() method for now. This is my own personal pet project so there are no deadlines and no production environment concerns. At the very least it will crash less than the MUD codebase I've tinkered with in the past - it once segfaulted three times in a row before starting up and I hadn't even changed anything yet. Corner cutting goes against the grain for me so I think I'll just write it from scratch. Perfectionism strikes again.

I figured the best way to learn was to hold my breath and jump in with both feet. This approach has worked for me so far, so why not now? :)

No need to apologise for not releasing code - I certainly understand and the experience will be good for me anyway. Actually, given that you could release the code, I'm not sure I would look at it anyway - this program is as much about learning the ins and outs of network programming as it is about creating an interesting / useful project.