Operating System as Social Network

20 Years Before Facebook, a Time-sharing OS Created a Unique Social Network. We’re Still Just Catching Up.

In 1982, Stevens Institute of Technology made headlines when they announced that they would require all incoming freshmen to buy a personal computer, which would be incorporated into the curriculum. I was 12 years old that summer and had been smitten with computers ever since I laid hands on a TRS-80 Model 1 at the local Radio Shack two years prior and subsequently started spending all my afternoons there. As I sat at our kitchen table in suburban New Jersey and listened to the radio news about Stevens’ computer initiative, I knew it was the college I would attend.

By the time I arrived at Stevens in the fall of 1988, the new computer curriculum had gone through a full undergraduate cycle and the computer culture had fully bloomed. Computers were well integrated into the curriculum, and hardware had progressed from an Atari 800 to the DEC Pro-350 and Pro-380 to the machine I received, a 286-based AT&T 6312.

What was ultimately far more interesting, but received no headlines, was what Stevens had done to the campus: they wired every single room on campus, from classrooms to dorm rooms, with Ethernet.[1] The ubiquity and density of computing and networking on campus, and the way they shaped communication and interaction between everyone on campus, foreshadowed much of how we interact with technology and each other today.

I don’t think many details from that exciting moment have been written about — I’ve certainly struggled to find any — so I wanted to try to capture some of the aspects of social computing at that time.


Daily Life and the Tools of the Trade

While every student had a PC on his or her desk, the heart of the campus network was a three-node VAX/VMS cluster housed in the computer center, visible behind glass in all its massive glory.[2] Every member of faculty, administration, and the student body had a unique login and email address. For students, the login format indicated whether you were an undergraduate or graduate, your class year, your first initial and some or all of your last name — for example, I was u92_mmorriss.

Not a photo from Stevens, but this 11/780 is an example of one of the nodes in the Stevens VAXcluster.

Email was centralized on this cluster — you had to log onto the cluster to send or receive any email.[3] That may seem odd today, but the internet at that time was incredibly small. So small, in fact, that every computer that was attached to the internet was listed in a file that we would print out occasionally.

This need for users to “gather” on a small set of machines to email or to run large computing jobs gave rise to novel online social interactions. These student-to-student interactions were primarily triggered through two VMS commands: show users and the awkwardly named finger.

show users did as its name implies: it showed all the users currently connected to that particular node in the cluster. With an additional parameter you could also check the other nodes in the cluster — this was important because people tended to have a preferred node they like to connect to. (Mine was, somewhat inexplicably now, SITVXC — as opposed to SITVXA or SITVXB.)

show users let you see the full name of the people connected to the node, along with some additional information. One such bit of information was how long the user was idle. Another was what command they were running. (This was at a time when you could only run one “app” at a time.) Socially this gave a hint about their status — were they sending email? Editing a document? It was like a richer version of the presence indicators in Slack or Facebook Messenger.

The show users command also listed another interesting bit of information: the “process name”. This was a very small field that by default displayed your login name (eg, u92_mmorriss), but which could be customized to anything you could fit in 15 characters. Students discovered it was the perfect vehicle to broadcast their current situation: their mood, what they were studying or listening to or reading, shared private jokes, and so on. (If you think that Twitter’s 140 character limit drives pithy creativity, imagine lowering it to a tenth of that.) It was not uncommon for one’s process name message to change a dozen times over the course of an evening. Some people wrote scripts to update their process name automatically, for example to slowly cycle through a long message.

What made this “current status” so powerful was how ingrained it was in the campus culture. If you were on campus, and in your dorm room or at a computer in a lab, you were logged into the VAXcluster. If you were logged into the VAXcluster, you set your process name. If you didn’t set your process name, well, that said something about you too. Students reflexively checked who was online anytime they were near a computer. Within days of arriving on campus, this habit was ingrained in the majority of students.

Where the process name was used as the equivalent of a status message, something called a plan file was the equivalent of a home page on the web. It was simply a text file in your “home directory” and took no technical ability to create — so nearly everyone had one. What people put in there varied as widely as web pages do today. Some used it as a blog; some used it to publish personal essays. Others showed off their ASCII art ability; still others put in favorite quotes or song lyrics, contact information, and other bits of vanity. Most people cycled through all of these options. (Professors tended to use it for syllabi and office hour schedules.)

Random example of what type of information was display when you finger’d a user.

You could view someone’s plan file through the finger command — if you entered finger u92_mmorriss, for example, you would see whatever I was railing about at that particular moment. If you were curious about someone, whether you knew them from around campus or had just been intrigued by their process name when you show users’d, you’d finger them to find out more. Online life and offline life were completely blended. It was the original version of Facebook stalking, minus the photos.

It’s hard to believe now, but in those innocent days, finger would also display the last time someone connected and, most shockingly, how many unread email messages they had and the time of the last email received. The social implications of making this information public were enormous, particularly on weekends.

Unlike Facebook, however, all of this information was ephemeral. Status messages set in process names and screeds contained in plan files would come and go along with students’ whims. This gave rise to a form of self expression that was in-the-moment and more experimental — ideal for college. Whereas today’s ephemeral apps arose out of exhaustion with social-network-as-permanent-record, the ephemerality of process names and plan files came from physical limitations of storage and network speeds, as well as the absence of easy screenshotting.

Along with these broadcast-type communication methods, there were a variety of tools for private communication. You could phone someone — that was the same as instant messaging today, not voice calls. Or you could turn to email for more personal — or more formal — conversations. There was also a command that would let you send one line messages to someone else’s terminal, much like a text message today. Small group chats were done over phone; big group chats were less common but done through relay (Stevens was on BITnet, and students sometimes chatted with our neighbors at CUNY and other colleges). [4]

Everyone used a combination of all of these tools, all the time. For the entire four years of my undergraduate years at Stevens, daily life was dominated by this unnamed social network.


25 Years On

Without design and without intent, what began as an educational tool transformed this 3,000-person campus into college sponsored, if not implicitly mandated, social networking experiment.

No doubt each of the individual communication methods described here seems rather quaint from today’s perspective. But taken as a whole, along with the universal access we enjoyed, they formed a social platform that had no public equivalent until fairly recently.

This magical period at Stevens continues to influence my perspective on current social networking. Despite vast improvements in hardware and software, I think there were communication modalities and social dynamics in play then that have yet to be fully realized on our networks today, particularly around online-to-offline interactions. It will be fascinating to see how long the experiences of that golden age will shape the future of our communications.


Footnotes

[1] Specifically, ThinWire — 10BASE2. Ethernet cards for PCs were so expensive at that time that they weren’t included with the purchased computer — they were merely leased to students and had to be returned at graduation. Students who were living off-campus were provided modems.

[2] Along with the main three VAX nodes, there were a number of miniVaxen, microVaxen, and Vax stations. Unix was definitely underrepresented on campus. There was a 3B2 in the physics department, a couple of engineering labs with Sun-1s, some Sparc stations, and later, a lab full of NeXT cubes. There was also a large Apollo machine and a Symbolics Lisp machine in the math department, where I spent the majority of my time.

[3] Interestingly, a Stevens class of ’77 grad, Mark Crispin, went on to develop the IMAP protocol for email that is still widely used today.

[4] File transfer was also available and used on campus, but it was more common to transfer large files (or large number of files) over floppy disk, sneakernet style.


Thanks to my u92_ and u93_ friends who read drafts of this post.