RFC 135
Response to NWG/RFC 110
Response to NWG/RFC #110 π
(Conventions for Using an IBM 2741 Terminal as a User Console for Access to Network Server Hosts) π
Okay bestie, the title block first π¦ This one's from the Network Working Group and it was written by W. Hathaway at AMES. It's RFC 135 (NIC 6712), dated 29 April 1971, and it updates RFC #110 β¨
Hathaway wants to swap out the conventions from RFC #110 for new ones π ββοΈ ngl the originals didn't really think about the correspondence 2741, they were kinda inconsistent, and they were a pain to use π Quick heads-up: the 2741 with the correspondence keyboard doesn't have every standard character. The missing ones are:
less than <
greater than >
logical not [1]
vertical bar |
So we must not use any of these characters in our conventions if we want the correspondence 2741 to work π
Dedicating a character to a special function is a real trade-off π You get the convenience of one key for the function, but you lose the ability to type that character as plain data, which means you need two keys. Hathaway thinks only two of the special functions from RFC #110 deserve their own key: "character escape" and "character delete" π
For "character escape," the cent sign [2] is the pick. It's on both the regular and correspondence 2741 keyboards, and it isn't in the ASCII set, so it's perfect π For "character delete," it's the backspace key, obviously π Yes, there's some need to type "backspace" as data (like for underlining output, etc.), but Hathaway says the trade-off makes it way better to give "character delete" a single key than to give "backspace" one and make "character delete" two keys π₯
Everything else gets a two-key combo: "character escape" [2] plus a key that names the function. Here's the lineup:
character escape [2]
character delete backspace
system delete [2]$
line delete [2]# or [2]#NL
logical line end [2];
line continuation [2]NL
ASCII control [2]@
About "line delete": the option of typing NL right after it lets you line up your margins without creating a null line. If you want a null line after a "line delete," you'd need two NL characters instead πΈ
Now the two new functions, "line continuation" and "ASCII control." π§ "Line continuation" lets you type a logical line that's too long for the 2741 carriage (or your margin setup) by spreading it across several physical lines. It's basically the opposite of "logical line end," which is so iconic π¦
"ASCII control" needs a little backstory π A bunch of ASCII characters are made by holding "control" and pressing another key. The "character escape" function can handle those too: "character escape" followed by a letter equals the ASCII "control" plus that letter, written as Xc (where X is the letter). So users just type "[2]A" wherever they'd normally type Ac, and it becomes super simple π
Some control characters need "control" and "shift" along with a letter, though (ESC is SHIFT Mc, for example). Also, someone who isn't a Teletype user might know they want a LF but not that it's a Jc. So "ASCII control" lets you type any ASCII control character as "[2]@" followed by its standard two- or three-character abbreviation π«Ά "Escape" becomes [2]@ESC, "line feed" becomes [2]@LF, and so on. Using variable-length abbreviations doesn't create any ambiguity, but Hathaway notes that implementers might find it easier to stick with the two-character abbreviations proposed in RFC #110 β¨
Last up: there are eight ASCII graphics that don't show up on either 2741, plus the "cent sign" and "backspace" themselves (without their special functions). Hathaway's mappings for those are:
[2]( for [
[2]) for ]
[2]6 for {
[2]9 for }
[2]/ for \
[2]' for '
[2]" for ^
[2]- for ~
[2][2] for [2]
[2]backspace for backspace
And finally, "<" and ">" aren't on the correspondence 2741, so they should not be used π ββοΈπ
Endnotes π
- [1] logical not
- [2] cent sign
[This RFC was put into machine readable form for entry] [into the online RFC archives by Lorrie Shiota, 10/02] π
π¦ where this page comes from
- ietf: rfc135 1971-04
- π girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms