RFC 93

Initial Connection Protocol 🎀

okay bestie, this is RFC 93!! 💖 It's by A. McKenzie from BBN, dated January 1971. It's a Network Working Group document, and it updates RFC #66 and RFC #80 🧚 (so it's basically a patch note for those two). Its whole job is to fix a little confusion in the Initial Connection Protocol (ICP) ✨

The ICP was first described in RFC #66 and then restated in RFC #80. While reviewing it, someone spotted an ambiguity 👀 It comes from the Host-Host Protocol Document No. 1, and it's about the message sent over the connection from "Server socket #1". Both RFCs say the message is "exactly an even 32 bit number" 🔢 But it's not clear whether that 32 bit number is supposed to come right after an eight-bit "message data type" code. The Host-Host Protocol has room for those codes, but it doesn't seem to absolutely demand them 🤔 which is so frustrating, honestly 😭

Only one ICP implementation has been written up in the NWG literature, and that's the one at UCSB (RFC #74) 💅 The UCSB folks read the Host-Host Protocol as requiring a message data type code, so they send a code of zero. Lowkey iconic of them 🫶

Steve Crocker said (in a private communication) that the Host-Host Protocol was meant to require a message data type code 🎯 So the recommendation is to amend RFCs 66 and 80 so they show the "even 32 bit number" preceded by a message data type code of zero 🌸 Zero is the only code defined so far, so that's the one 🧚

Main character energy: this patch exists so future implementers don't have to guess 🦋

[ This RFC was put into machine readable form for entry ] [ into the online RFC archives by James Thompson 4/97 ] 💜