That guide is somewhat recommended by many college courses on Unix network programming. That's why I gave you this link.

1. Really? When I read it some time ago I didn't recall I saw this problem. Perhaps a few places may be. I may help clarify if you give me exactly where the omissions go. By the way, they're functions, not methods. When you talked about methods I did take some time wondering what you meant.

2. listen() lasts until you teardown the socket.

The chapter describing the functions do not have full examples. On the next chapter (Client-Server Background) you'll find examples in full that you can try on your system.

3. Assuming you're using TCP (not datagrams). Once a server accept() a connection, from the accept(2) manpage, "it creates a new connected socket with mostly the same properties as s, and allocates a new file descriptor for the socket, which is returned." You can then use recv() on this returned socket (not the one which you called listen() on!) to receive data. You'll need to allocate a buffer, and very likely you should put it in a loop to get the incoming data in chunks. Somewhere in the document it mentioned encapsulating your data by putting a header indicating the length of data. You can use this scheme to decide whether you would like to make a dynamic buffer which fits all the data, for instance so you don't need multiple recvs.

4. Theoretically you can, by using a cross-compiler, then one can compile a Unix C program on Windows or vice versa. I don't have any experience on gcc cross-compilers though, so you may need a search on google for this.