Among the many flaws in this article is the fact that it cites "first login request". The history of telecommunications covers three fundamentally different types:
- human-to-human
- human-to-computer
- computer-to-computer
- human-to-human
- human-to-computer
- computer-to-computer
Telecommunications starting around 1830 was primarily human-to-human. People sent telegraphs to each other. People made phone calls to each other. It was a means for humans to talk to each other.
Over time, humans got removed from the loop. Early teletype machines from around 1910 where mechanical devices that could send and receive text without direct human interaction, but it was still text intended for humans.
When computers arrived, the system used for human-to-human interaction was changed to human-to-computer. Those same teletypes were now repurposed to type commands into the computer, and to receive printouts from the computer.
The early ARPANET just extended this interaction. Instead of requiring a direct long distance link between your terminal (human) and computer, you could share long distance links.
ARPANET wasn't unique in wanting this sort of design. The telcos were building this sort of service. IBM was also offering this sort of service for their mainframes. Evolving human-to-computer (terminal-to-mainframes) was something everyone was doing.
The defining feature of the Internet, though, is computer-to-computer interaction. This is a vastly different problem. It's why you can't really look at early computer networks like IBM's or even the ARPANET as progenitors of the Internet.
You see this in early designs of TCP/IP. What the designers were most familiar with was human-to-computer interaction. That's why FTP, the "file transfer protocol", is defined as running on top of Telnet, a human-to-computer protocol.
By 1980, though, we had the "personal computer" revolution. People stopped using "terminals" attached to "computers", but had the computer itself on their desktop. Before 1980, the "computer" was in a different room, a different building, or even a different state.
This broke the design of the 1970s, which was primarily about terminal-to-computer networks. Everyone had to pivot and redesign everything to be computer-to-computer networks.
As I've tweeted about before, you are taught the "Presentation Layer" as layer 6 of the OSI model, but you never really understood it. That's because it was designed for human-to-computer interaction of the 1970s and was already out-of-date by the 1980s.
Xerox was at the forefront of personal computers in the 1970s, and already had computer-to-computer protocols designed when Internet engineers came around. Much like how Jobs based the Macintosh on Xerox designs, academics leveraged Xerox designs for the Internet.
When BBN added a TCP/IP stack to BSD 4.2 in 1983, they also added a Xerox XNS stack as well. Because these were the two protocols that could support computer-to-computer interaction, with XNS predating TCP/IP.
Throughout the 1980s, XNS (in the form of Novel IPX) carried vastly more network traffic than TCP/IP. It's really an accident of history that TCP/IP became the eventual standard rather than XNS.
The Internet as we know it really dates from 1983. What you choose before that date to call early Internet stuff is largely arbitrary. You could focus on RAND reports, or early ARPANET stuff, Xerox XNS work, or AT&T telcoms networks, or a lot of other things.
