How and what did it mean to connect to ARPANET from home?

Question

In the iWoz biography, I read:

Now, TV terminals I already knew a little about. […] I'd already built a terminal that could access the ARPANET, the government-owned network of computers that was the predecessor to the Internet. […] Teletype systems cost thousands of dollars […] but I built a system using a Sears TV and a cheap $60 typewriter keyboard.

Now I had imagined the ARPANET was something of an early precursor of TCP/IP-like capabilities, running on a physical network between a handful of universities and/or military bases. How would Steve Wozniak have connected to this with a rudimentary terminal from his home?

Surely he would have connected via some sort of IMP minicomputer gateway, and presumedly dialing into that using a serial modem of some sort? But then what? Would Woz have serialized/deserialized raw packets manually, and typed them in hex as it is said he later did on-the-fly with 6502 machine code? That doesn't sound very fun.

I've noticed this biography can be rather sloppy with details, so perhaps he "merely" connected to a terminal host system that was also connected to ARPANET? But even so, what would this have entailed? Would he have "shelled into" some sort of pre-UNIX server and been able to run a variety of ARPANET-aware programs from that host? Or would it have been more like a BBS where a single program would "answer the phone" and offer a small menu of various options in this case related to/backed by a wider-area network connection?

What could a "home user" do on the ARPANET in the months/years leading up to the founding of the Homebrew Computer Club in March 1975?

Answer 1

You dial in with a terminal to a TIP (Terminal Interface Processor), which then offers a simple command interface to connect you to some host machine.

Living Internet web site

Wikipedia

Functionally, this is no different from late 20th-century terminal servers. The terminal connects to a device (a small dedicated computer, network attached) which implements (1) the network protocols needed to carry terminal traffic to a remote host, and (2) a local command interface for indicating what host to connect to.

This TIP user's guide from an Arpanet conference shows the procedures for accessing remote hosts.

From memory, the TIP may or may not be integrated into the IMP.

Answer 2

Simply, he built a serial terminal which reads ASCII text from an RS-232 connection and displays it on the screen. Then when you type something on the keyboard, it sends that back over the RS-232 connection, character for character. There's no TCP/IP packets involved, just straight ASCII.

Then, he connected that to a serial port on a computer connected to then ARPANET that provides a shell over serial, or he hooked up a modem to his terminal and dialed in.

He didn't need to decode raw packets because it wasn't a SLIP connection, just a shell. He could type things like ping or sendmail or ftp and interactively issue commands, or telnet to a different computer and do it from there.

Answer 3

re: "What could a "home user" do on the ARPANET in the months/years leading up to the founding of the Homebrew Computer Club in March 1975?"

Kids in the mid 70s whose parents had terminals and modems at home could try to find access phone numbers (via word of mouth) that didn't require a password, and once connected, telnet around looking for systems they could access. Online activities were mostly text-based games and (what else?) chatting with random other users. The MIT AI lab system was quite permissive as I remember it.