by Darius Kazemi, April 22 2019
RFC-112 (PDF link due to important info missing from text version) is titled “User/Server Site Protocol: Network Host Questionnaire”. It's by Thomas O'Sullivan of Raytheon Data Systems and dated April 1, 1971.
The technical content
RFC-106 was sent out a month prior to this and was a questionnaire that asked different ARPANET sites to answer some questions about their computing setup. Specifically a committee is trying to come up with a more formal Telnet specification and having this information will help them add needed features to Telnet. This RFC summarizes the results of that survey, one month later.
The results are hand-written and scanned an a bit hard to read but my best reading of the results shows a few things of note.
First, there is no consensus on a common terminal format between the sites. TTY ASCII format is popular, but some sites (like Lincoln Lab and RAND) don't support it. A slim margin of the computers themselves use ASCII internally but there is a sizable minority of computers that use EBCDIC internally.
Of all the sites, only Raytheon performs error detection. Only three sites support punched paper tape input.
Interrupt mechanisms across systems are by far the least standardized. Almost every single site's computer system uses a different interrupt mechanism from all the others, though notably the early ARPANET sites running PDP-10 machines are actually pretty similar in this regard, where
ctrl+c breaks a program and returns you to the command shell. This is how most operating systems that people are familiar with today handle this kind of interrupt, which makes sense if you look at the history. The PDP-10 was explicitly copied (through the CP/M operating system) by MS-DOS and thus Windows. And the PDP-10's TOPS-20 operating system had a huge influence on Unix/POSIX, so it probably came from there as well.
Generally speaking, looking at this chart makes it clear that IBM 360 and Multics systems feel very foreign to me as a longtime POSIX operating system user, whereas the PDP-10 running TOPS-10 seems like it's doing things “right”. Of course there's not a right or wrong here, but TOPS was a clear winner in the battle for how things ought to be done.
Although this RFC is dated April 1, it is not an April Fool's RFC, which is a tradition that began in 1978 with RFC-748, though there are some humorous RFCs that predate that one. More on this here.
You can request a login from The Living Computer Museum in Seattle if you'd like to try remotely connecting to a TOPS-10 system and playing around!