mca blog

a recent thread on the atom list drifted into what status codes are appropriate in response to the DELETE method. after a handful of back-and-forth posts, roy responded with an excellent observation:

"Honestly, the number of status codes defined for HTTP is so small that it takes less time to write an implementation table for every single code than it takes to have this conversation."

once you swallow the pill that HTTP is all about building a robust state machine (server) and intelligent client (client browser), then comments like this make perfect sense. i have to keep reminding myself of this fact - almost every day!