by Brian Dear
—
Nineteen seventy-three was a busy year for the rapidly evolving world of computer conferencing systems, primitive and crude as they might be, not only at the University of Illinois but at a handful of other campuses and research labs around the country. In addition to Discuss (the simple message forum that had gotten Stuart Umpleby in so much trouble), there were other places on PLATO to post messages and share information with other users. One lesson, Pad, was, like Discuss, a simple note-and-response system where users could post short notes and then other users could reply. Pad was a hangout with no particular purpose other than posting silly messages, observations, jokes, complaints, and non sequiturs. Naturally, it was very popular.
Even plain TUTOR source code lesson files themselves were used for communication. A whole series of lesson files, named “notes1” through “notes19,” were simply source code files where instead of source code, users posted messages. But as everyone would discover, using a source code file for group communications was untenable and, in the long term, unmanageable. For one thing, it meant that only one person could edit the file at a time. As usage scaled, it became more likely that someone else was editing the file when you wanted to post something—so you had to sit and wait. “Reading and writing notes in this manner became a problem as the system grew,” says Kim Mast. “There were limited cases of people destroying other people’s notes. While users were encouraged to write the note in their own file and copy it into the ‘notes’ file, every once in a while someone would forget, and be editing notes for an extensive period of time, locking others out.”
Any growing community using a time-shared computer that requires the honor system to facilitate the act of communicating between members of the community will eventually fail, because as the community itself grows, so grows both the number of people who won’t abide by the rules and the number of people who forget the rules. The result is that things get out of hand. Because the “notes” files were open to all users, anyone could edit anything, innocently or maliciously. They could change what someone else said, post something and sign it as having been written by someone else, or even delete not just a single posting but all of the postings. It was clear that a real program, an official system program supported by the CERL staff, was needed that offered all of the benefits of these simple “notes” files but none of the disadvantages. One day in the early summer of 1973, another mass-deletion incident occurred in the “notes” files, and for Paul Tenczar it was the last straw. He assigned David Woolley, a junior systems programmer, a new project: go do something about the “notes” problem. What emerged on August 7, 1973, was a new feature on PLATO called, simply, Notes. David Woolley was just seventeen at the time, a UI freshman. He was one of the new wave kids, and like several other junior programmers at CERL, he had come from Uni High.
Notes would become Woolley’s claim to fame. What started as a quick solution to an urgent problem would become one of the foundations of the PLATO online community, and would keep him busy for the next several years.
Woolley settled on a simple organizing schema for messages. A “note,” which might today be called a “thread” or a “topic,” consisted of a “base note,” and one or more “responses,” stored and read linearly in chronological order. The base note was the message left by whoever started the new note. The responses were messages left in reply to the base note (or in reply to subsequent responses). “I came up with a design that allowed up to sixty-three responses per note, and displayed each response by itself on a separate screen,” says Woolley. “Responses were chained together in sequence after a note, so that each note could become the starting point of an ongoing conversation.” Base notes and responses were both limited to up to twenty lines of text. If a user needed to be more verbose than twenty lines, they needed to enter the first twenty, then “reply” to their own text, and enter more text. But they’d better hurry, because someone else might—and often did—reply first.
That was it. Nothing fancy. But its simplicity made it understandable, and that enabled the community to embrace it enthusiastically. Woolley also provided a simple notes index display showing the list of most recent notes, how many responses there were, if any, and the date that the base note was written. In keeping with the severe constraints brought on by disk space and screen space, base note titles were kept brief, just fifteen characters. The index display pages listed fifteen base notes at a time, along with the base note number, date it was written, title, and number of responses. The NEXT key paged forward, BACK paged backward. The result went far beyond the primitive notes1 through notes19 source code file editing the community had been relegated to before.
Tenczar would have been happy with a solution that allowed for a single response to a note, so that a systems programmer could indicate that the problem had been fixed or not. But Tenczar was considering only one use case: using Notes as it had been used among the systems staff to receive reports of system problems from the user community and then reply back with quick one-line responses. Woolley realized that a fully functional Notes program ought not be restricted to this one narrow “customer support” use case; rather, it should support conversations of any type.
On August 7, Paul Tenczar posted the first note in the new “System Announcements” notesfile:
newnotes Note 1
8/7/73 11:07 pm CST pjt / s
Since you got here, you will undoubtedly note that we now have a new system of user/system notes. We hope that they will greatly speed up your browsing…and provide us much greater protection from note-destroyers!
Please direct any comments about these new notes to Dave Woolley.
Old notes are obtainable by editing files -notes1- through-notes19-.
The first release of the Notes program supported three categories: System Announcements, Help Notes, and General Notes.
“System Announcements” or =announce= (the traditional way PLATO users mentioned notesfile filenames online was to surround the name with equals signs) was the official word on new features, problems with the system, changes to the way things worked, and other news. Given that TUTOR was constantly changing (either one of the attractions of the language, or one of its biggest drawbacks, depending on who you spoke to), keeping up with the =announce= notesfile was crucial. A tradition developed fairly quickly that the person on the systems staff who actually did the work to create a new feature would get to post the news about the new feature.
“Help Notes” or =helpnotes= started out mainly as a place to get help related to PLATO but would become a fantastic resource to the community seeking help on pretty much anything: Need help programming something on PLATO? Trying to find a reputable automobile mechanic in Champaign-Urbana? Stumped as to which flea collar might work best for your dog? “Help Notes” was the place to ask your question. Oftentimes you’d receive a reply within minutes if not seconds.
“General Notes” was for everything else, and became known as “Public Notes” or =pbnotes=, for messages to and from the entire community.
Over the next three years, Woolley would continue to add features to the Notes program. By far the most notable change occurred in the early winter of 1976, when Woolley announced to the world that he was expanding Notes so that there would no longer be just three “sections” of the program, one for system announcements, one for help notes, and a general notes repository, but, instead, the program was being redesigned so that there could be any number of notesfiles, on any subject imaginable. The Notes program would become the engine that managed and presented these notesfiles, but there could be, and soon would be, thousands of notesfiles, each dedicated to a specific subject. A tiny sampling of some of the hundreds of public notesfiles that subsequently popped up included:
abortion aerobics anim
animbest animations announce
antic apartments apple
astronotes audiofile avatrade
bald believers bereans
bestwishes bicycle
books
bowlers brandx calcnotes
cars cerlnotes chessnote
comix commodor consumer
donkey dreams drugnotes
ednotes eenotes elephant
empirenote events filmnotes
gamesnotes garden goldwing
hebrewnote helpnotes ibmnotes
ipr jobnotes jokenotes
karate kidnotes lawnotes
lunch magicnotes medieval
micronotes musicnotes nature
pad pbnotes petnotes
philosophy photonotes platopast
poets recipes relativity
romance sayings science
scifi scouts scuba
sexnotes shuttlen skinotes
skydive soapnotes spacenotes
sportnotes startrekn strange
tennis theology trains
trivial tvnotes unixnotes
videog wantad whonotes
wilderness wine wishnotes
One pair of notesfiles grew in notoriety over the years: =derfnotes= and =derfbest= (the latter being “the best of =derfnotes=”). =Derfnotes= was a repository of derf messages. A derf (backward for “Fred”) was a note written using someone’s signon without their permission, usually because that person had gotten up and walked away from the terminal. To derf someone was to write a derf note, which might range from a quick “I are a derf” (often that is all the time one had to type before the derfed individual returned to the terminal) to long embarrassing essays on the risks of leaving one’s terminal unattended. The more well known the person, the more tempting it was for someone to derf them.
While Dave Woolley worked for all of his undergraduate years improving and expanding the functionality of the Notes program, his old Uni High classmate Kim Mast had been assigned a parallel project, Personal Notes. Personal Notes was to be PLATO’s email system. It worked very similarly to the Notes program: users would use the same text editor to compose a message, and similar keyboard commands to navigate among the messages. Personal Notes, or pnotes (pronounced p-notes), as they came to be called, was launched August 10, 1974. Now PLATO users, at least users with “author” and “instructor” signons (as opposed to mere “student” signons, who were using PLATO strictly for its instruction), could “email” each other.
One thing that particularly made Notes and Personal Notes successful is their exploitation of a subtle but hugely important aspect of the PLATO system’s unique computer architecture. Thanks to the Fast Round Trip, the Notes and Personal Notes programs provided additional responsiveness to users. Even though a note’s twenty lines of text might take a few seconds to plot on the screen, given the 1,200-bits-per-second bandwidth constraints of the time (agonizingly slow by today’s standards), all the while the system was monitoring furiously for the slightest keypress from the user. The resulting user experience was quite something to behold compared to typical non-PLATO computer programs of the era. On other computers with display terminals, if you requested a display, a block of text on the screen, it was often very hard to interrupt and stop the display with any degree of accuracy or control. Oftentimes you had to press a certain key sequence, like Control-C or Control-S, multiple times to get the text to stop. And then who knows where you were in the program or whether the program would respond to any other input. Whereas on PLATO, as information began pouring onto the screen and within the first few words you realized this was not what you wanted to read, or that you’d seen it before, there were a variety of single-key keypresses that would immediately stop the display midstream, and go immediately to wherever it was you wanted to go. For example, if you pressed the LAB key while the text of a base note was displaying, PLATO immediately stopped, the screen instantly erased, and the text of the first response began to appear. Likewise, you could press LAB again, and jump to the second response. Then press BACK and jump back to the base note. Press the “9” key and it would jump nine responses forward. Press the minus (“-”) key from there, and the system showed you the previous response, number 8. You could press a long sequence of keys in succession, staying ahead of the display but knowing exactly where you were, for example, LAB, LAB, BACK, 9, 9, 9, -, and PLATO was right there with you, stopping text from displaying the moment you pressed a key, and landing you on the twenty-sixth response to the base note. The entire PLATO user experience was this way: working at the speed of thought.
When a long list of notesfiles continued getting longer, the user community complained. Not because of the sheer number of notesfiles, but because of how hard it was to a) remember them all and b) keep up with the conversations going on in them. You might visit a popular notesfile and stare at the list of notes: perhaps fifty new ones in the past few days since you last visited, and each of those fifty have maybe twenty-five or more new responses. People were overwhelmed. Where to start? The success of PLATO Notes brought on a happily solvable problem.
Dave Woolley added a DATA key option that enabled users to go through notes and responses chronologically. Another systems programmer released a special -jumpout- feature that enabled PLATO authors to write their own programs that took advantage of a “cycler” tool that would roll through a given list of notesfiles and only show you what you had not already read. (The World Wide Web would respond to a similar need as the number of websites, news sites, and blogs exploded in the late 1990s and early 2000s, by offering RSS—“Really Simple Syndication”—feeds that kept track of what you had read and what was new.) Unfortunately, due to a poor implementation of the “cycler” function and overuse by enthusiastic authors, PLATO could not handle the load put on the mainframe hardware. Rick Blomme then directed John Matheny, another CERL systems programmer, to create a centralized, more efficiently designed, system-supported utility, which got the name “Notesfile Sequencer.” It was an enormous jump forward—another catalyst that not only accelerated a PLATO user’s productivity, saving them enormous amounts of time, but in a way contributed to the general “acceleration” of PLATO users themselves. As the sheer amount of information and conversations kept growing, users could not keep up, and needed new tools to help them cope. With the Sequencer, users could create a personal list of favorite, must-read notesfiles, be it five or five hundred long, and the Sequencer would then automatically step through every single notesfile and only show the user those notes and responses the user had not yet seen.
—
Around the time that Dave Woolley originally built Notes, in August 1973, Karl Zinn, a professor at the University of Michigan with a longtime interest in computer-assisted instruction, had taken on a project with NSF funding to build an online forum. Zinn was active in the CAI community, knew everyone, went to the conferences, and published research. He knew Bitzer well, and visited CERL numerous times. He had seen PLATO Notes and related his impressions to a graduate student he’d hired named Bob Parnes. Zinn and Parnes called their system CONFER. “I showed Bob these various things, and wanted to borrow ideas from other places, or get somebody to agree that we could adapt something and create it for ourselves. So I would have been planting ideas with him and I influenced—every time he came back to me and said he was doing it this way or that way, I would say we really need more of this or that, because I was thinking of classroom use and he was still thinking of committee use at the time.” CONFER wound up having mechanisms for voting on issues, but was also organized in a “comb” structure like Notes had been. A “comb” conferencing system is one with individual discussion files, each of which can contain individual “notes” or threads, each of which can contain a number of responses. This was how PLATO Notes was arranged. There was one other conferencing system that also inherited this idea from Notes, called PicoSpan, which was a commercialized version of CONFER written by a developer named Marcus Watts, who along with Larry Brilliant created a company called NETI to market and sell PicoSpan. Brilliant was a friend of Stewart Brand’s, creator of the Whole Earth Catalog, and around 1984 Brand and Brilliant started working a conferenc
ing bulletin board service running on PicoSpan called the Whole Earth ’Lectronic Link—The WELL—a service that launched in 1985. Computer histories, even recent bestselling ones, often cite The WELL as the first online community with a conferencing system. While PLATO Notes came out a dozen years earlier and had a larger set of users around the world, the two communities that emerged on these two platforms, PLATO and The WELL, were remarkably similar in how they conducted themselves, no doubt influenced by the “comb” architecture of forums, threads, and linear responses.
—
It is important to appreciate just how far ahead PLATO’s suite of communications capabilities was in 1973–1974, and would stay that way for a solid ten years. Non-PLATO users had no point of reference to understand what PLATO users did or why they were so excited about working on the system. The sense of living an accelerated “digital life” was something only an actual PLATO user could understand. One way to illustrate the point is to consider a typical 1970s family dinner conversation. Family members might talk about, say, what they had done that day or what was in the news. If one family member was a PLATO user, they might casually mention that they talked with people in Hawaii, Delaware, New York, Florida, Illinois, and other places all that day. Other family members might react with horror at the cost of the long-distance phone bills for all those calls, until the PLATO-using family member pointed out that, no, these were TERM-talks. “TERM-what?” would be the typical response. The rest of the dinner might be spent attempting to explain to the puzzled, increasingly concerned family members what exactly TERM-talk, Talkomatic, pnotes, notesfiles, TERM-consult, and the Notesfile Sequencer meant. Most likely, family members would have fled the table long before the explanation was over.
Contrast this with a hypothetical family today: if they’re gathered around a single table at all, they probably all have a fork in one hand and a smartphone or tablet in the other, each family member lost in some private interaction on their personal screens, oblivious to the cares, concerns, or interests of other family members, everyone doing largely the equivalent of TERM-talk, Talkomatic, notes, pnotes, the Sequencer, and the rest.
