Abhay Saxena created code that was both easy to read and easy to understand. He hid the fixed size read buffers behind a good interface, thus reducing the complexity of the rest of his code. The solution was concise and correct and easy to verify as such. His code can be found in /mit/6.033/lab/examples/lab1/ark3.
Patrick McCormick had another good solution. Instead of going just a character at a time, he decided to read the input one line at a time. His code can be found in /mit/6.033/lab/examples/lab1/pmccormi.
Apache is a freeware, industrial strength web server with full source code on-line. If you're wondering how it's done in the "real world", this is as close as it gets.