Have you ever used Usenet before the Eternal September?

Post Reply New Topic

Since the damned cloudflare is not letting me edit the above post, I'll post it here.

I used to regularly take a DEC VT-100 terminal and a modem home from work over a weekend so that I could dial into the PDP-11/70 computer from home and work.

And here this whole time I thought that a dumb terminal was intrinsically reliant on a computer in order to do anything useful.

Edenthiel, Ah yes, ATDTnnnn Pretty soon we'll be reminiscing about JCL on paper cards.

MrBigmouth, The dumb terminal DOES depend on another computer - you're right. It's just that the other computer is on the other end of the phone line - like the mainframe at work, or the BBS or Compuserve.

A modem was not a requirement to connect a dumb terminal to a computer. It was a requirement if you wanted to connect a dumb terminal to a computer over a telephone line.

While I have a computer on my desk at the office, my main device for many tasks is a VT100 compatible dumb terminal connected to a unix machine via an RS-232 cable. Because no telephone line is needed, a modem is not necessary.

When I worked at various companies in the 1980s, we used VT100 terminals connected to a PDP-11 computer using RS-232 cables. The modem was only necessary when connecting to a computer via telephone line.

For what it's worth, I wrote a terminal emulator in assembly language back then for a PDP-11 running the RSTS/E operating system.

The program you ran to invoke the terminal emulator was either the shortest program ever written or tied for the shortest program ever written -- since it was zero bytes long, you couldn't write anything shorter. RSTS/E was built around a run-time system that would handle your program. I wrote the terminal emulator as a run-time system. You would then have to run a program to invoke the run-time system but the run-time system to use to execute the program was set by an attribute on the file. So I created a file of zero length and set it to choose the terminal emulator run time system.

The terminal emulator didn't care about the program used to invoke the terminal emulator so it could have been anything you wanted. For example, I could have had a gif of a Playboy foldout and set it to invoke the terminal emulator and it would have done so.

I thought about using the file to contain some initial commands to use when connecting to the terminal emulator, but I liked the idea of having written the world's shortest useful program more.

Anyway, with the terminal emulator, the modem was attached via an RS-232 cable to the PDP-11, not to the VT100 terminal.

eric76, good point about a modem not being necessary. Also, that is a wonderful story! I love that it was indeed a zero-byte program that did actual work.

In an old barn on some property we have upstate sits a mini with no discrete CPU (S-100 maybe?) but with something bigger than a Shugart SA1000 - and a single dumb terminal connected as you describe. It's also how the Alpha Micros used to be connected, too, although at that point it was just as likely the user in a bank or dealership would be using a dedicated $1200 PC to run a terminal emulator. IIR, it was just a simple tx/rx/gnd connection.

The technique is also still used extensively today for small (and not so small!) embedded systems such as routers. I think I have three different Cisco serial cables sitting on my workbench, including one with a rs-232 to usb converter. And most SATA drives still have a set of serial pins, very useful for resetting the drive when it has buggy firmware.

The program itself was 0 bites long. The attribute was a field in the directory system and not part of the file.

On the PDP-11 with RSTS/E, whenever you executed a program, you had to invoke a run time system which oversaw the execution of the program. If it was an executable binary such as with Assembly Language or Fortran, then the run time system would load the program into memory and then execute that program. If it was interpreted such as BASIC, it would load the program into memory and interpret the code.

In this case, it was the most trivial program of all -- zero bites. All the program needed to do was to invoke the operating system. I could have written it to execute any commands from the program as if they were typed with the keyboard and display the results on the monitor. If there had been no commands, once again zero length, then it would have just done what was typed at the keyboard.

But yes, a zero length program would still take up a directory entry in the file system. If not, then it couldn't be found and executed.

https://youtu.be/_COP3cyN7zg?t=18


All of this computer nostalgia made me think of this for some reason. Lol!