Nestled in a cozy office complex in Cambridge, Massachusetts, due west of Harvard and the Massachusetts Institute of Technology, is the new corporate headquarters of the legendary Underground Empire, Infocom. Perhaps embassy is a more accurate description of the facility, as there is certainly the feeling of stepping onto foreign, even alien, soil. Impressions of Infocom's several stories overlay and intertwine with the threads being woven on the loom of reality. Approaching the entrance, we recall the promise of a special surprise at the end of our visit, and that, in true Deadline tradition, if we keep our eyes open, we might discern clues as to what that surprise is. Come along and see if you can anticipate the secret with us.
Upon entering, you find a strangely familiar lamp sitting on the receptionist's desk. As you instinctively take it along, you notice an attractive paperweight: a sword embedded in a crystal rock. A man resembling a prospector sits waiting in the lobby, and the sword glows brightly blue for a moment as we enter the offices. After wending through a series of winding, twisting passages, we arrive at the Wizard's Den. There is the Wizard of Frobozz himself, bent over, gazing intently into a black sphere.
Marc Blank looks up from the Apple and smiles.
Before the Fall. To understand Infocom is to return to the pre-Apple days of the early seventies at MIT. As Joel Berez, president of Infocom, relates:
"We all at one time or another had worked for a research group that's now part of the Laboratory for Computer Science at MIT. This group was strongly result oriented. It was called The Dynamic Modeling System; now it's the Programming Technology Division.
"We believed it was important to develop programming tools before trying to develop applications. Over the years we had assembled an extremely advanced set of tools. Marc Blank lived next door to me when we were undergraduates and we both worked in this group.
"A dream of the leader of the group, Al Vezza, was that someday he would bring together all the people who'd been involved with the group to start a commercial venture using the same techniques that had been so highly successful. Of the ten founders of Infocom, eight came from the group.
"We also have three generations of teacher/students here; Dr. Licklider, in a sense, was Al's teacher, and Al was ours."
Joel Berez sets down the translator, as Marc Blank materializes to introduce us to one of the cocreators of Zork, the Great Implementor himself, Dave Lebling.
Moving through a door to the west brings us into a large conference room, where we sit around a shimmering table; everyone avoids touching the table. Water seems to glisten in an African painting on the wall where crocodiles and hippos swim serenely. You notice, though, that the green bubble on the dam upstream is not flashing. The robot in the corner puts down its apple and, while whistling, serves tea. Strange, your cup has a slight discoloration in the bottom.
The Underground Empire existed long before Infocom's founding, so Lebling and Blank relate how they first opened the door to Zork; the background music plays an oldie by the Taylor Mills Road group.
As Lebling explains it, "We had so much fun playing Adventure that when it was all over, we said, 'Oh dear, there are no more of these; how awful!' "
Blank, officially vice president in charge of product development, adds that there were some areas where they thought they could improve upon the original Adventure, which had been written in FORTRAN.
"It really bothered us that if you said 'Take bird' it would put the bird in the cage for you--sort of doing things behind your back. I suppose it's the nature of computer programmers everywhere to want to outdo the other person, to be a little better, especially in games.
"There were primarily three of us who worked on the original Zork: Tim Anderson, Dave, and I. In 1976, Tim and I started with an eight-room dungeon on a PDP-10, written in Muddle, or MDL, which was developed at MIT in the early seventies as a spinoff of Lisp.
"Then we began thinking of problems to overcome for the adventurer. The silliest thing in Adventure was that you always had to say, 'Take ax, throw ax; take ax, throw ax.' So we decided that anything you threw at Zork's troll, it would eat! A lot of things in Zork pay homage to Adventure."
Lebling and Anderson's Inferno. "We were thinking all along about an underground world. Tim and I had each read a lot of fantasy and it struck us as picturesque. We liked the idea that this was going to be like a book, like a story, so the player can say, 'I can see this world.' So we spent a lot of time on developing descriptions and trying to make the world seem more realistic."
That's where Lebling comes in. "I'm in charge of the purple prose; and I did the geographical design of Zork I."
Blank explains how Zork came to be Zork. "We'd generated this game with eighty places and five problems, and we didn't have a name for it. We just put it on the system. We had to have some type of name for it--files need names. Zork was a nonsense word used around the lab at the time as an exclamation--like, 'Zork, look at that!' The name stuck.
"We then added direct and indirect objects so we could tell it things like 'Kill troll with knife.' The parser got more and more sophisticated. As things progressed, we understood prepositions, we used adjectives, and we differentiated between objects with adjectives."
By June 1977, the first version of the mainframe Zork was finished. Blank had done the bulk of the work with help from Dave, Tim, and Bruce Daniels. Daniels is now working for Apple. He had done the coal mine section in Zork I and helped on the puzzle in Zork III. During 1977 and 1978, the areas of Zork II and part of Zork III were added. That is as far as Zork was developed on the mainframes.
How had Blank found time in his hectic, pressure-filled schedule at MIT to do all this massive work on Zork? Says Blank, "Oh, no. I had graduated from MIT in 1975 with a degree in biology. At the time of Zork's development, I was in medical school at Albert Einstein in the Bronx in New York City." Ah. That explains it. "I graduated from there in 1979. So how did I write Zork, in Boston, from New York? Easy. I used to drive up every weekend and work with Tim on it. During the week I would think of ideas to try out. I like biology and was sort of interested in medicine, but I love playing with computers, and I love games and programming." Enter Infocom, stage right.
On Goldfish Pond. Al Vezza is the Prime Mover of Infocom and chairman of the board. He still works at MIT full time, where he holds the impressive title of senior scientist and is currently acting associate director of the lab. Gazing out a rear window at the aspen trees and the goldfish pond, Vezza recalls, "Our main objective when we started this company in 1979 was to use our technology to sell products. At that time we didn't have a specific product in mind, but my firm belief is that the way you create new products, with new ideas, is to get a bunch of bright people, put them in an environment that is stimulating, and let them go at it."
Over the summer of 1979, Berez told Blank about Vezza's exciting new venture. Blank suggested that Zork was certainly marketable, and, according to Joel, "Marc and I kicked around the idea of putting Zork out commercially; we designed a machine-independent language for games that summer." Berez, at that time, was in the Sloane Management School at MIT earning his business degree. They took the idea to Vezza, who agreed to let Zork be Infocom's first venture.
From the summer of 1979 to the spring of 1980, Blank wrote the compiler and Lebling wrote the assembler for the new language. Bruce Daniels wrote the Apple Kernel. During this time, Zork for the micro was being developed independent of MIT on rented space on a DEC-20 from Digital Equipment.
In the spring of 1980, the decision to license Zork to Personal Software was made. Vezza relates why it was not a hard sale.
"We talked to Dan Fylstra from Personal Software about Zork. Turned out that when Dan had been in business school in Boston, he had played Zork on my computer while doing a thesis. So he knew what Zork was all about."
The Wizard and the Shape Changers. Zork did quite well under Personal's umbrella, but in the fall of 1981, Personal underwent a corporate metamorphosis into VisiCorp. The new image involved VisiCorp's dropping all its game lines and concentrating on business programs, so Infocom took back the rights to market the Zork line. This created quite a problem for Infocom at first.
During a moment's pause in Blank's narration, you watch a crystal phoenix being reborn in the air above the shimmering table. Blank continues the story.
"We've been in business three and a half years, but only one year has involved marketing under our own name. Last fall we started an incredible scramble. We had Zork II ready, but we didn't know anything about packaging or advertising. So we got a terrific agency, Giardini/Russell in Watertown, Massachusetts. In two days they came up with the Zork logo and the manuals--really quick; and we had Zork II out for Christmas. We took our first office last September in the Faneuil Hall Marketplace in Boston. It was only one room and, in true Zork fashion, required a special key for unlocking the elevator to get up to that floor. Joel was the first full-time employee and started last July. I came on at the same time and became full time in January. We now have a staff of four people and we'll be expanding to ten within four months. Others, like Dave, have regular jobs elsewhere, working here evenings and weekends. We moved into these new offices in January."
As Blank shows us around the offices, we notice all of Infocom's package fronts framed on a wall, including the two new ones, Zork III and Starcross. Next to Starcross is an empty frame filled with ivy. Perhaps the ad agency is from nearby Harvard.
Great Caesar's Muddle! As we walk, Blank and Berez explain the Muddle language and why it's so central to Infocom's success. "Muddle, like us, is a result-oriented, function application language," says Berez. "A program is formatted, debugged, and compiled on a DEC-20 in Muddle, then translated into machine language, which is interpreted on the Apple. Thus programs are easily transportable between different machines. Currently, all our programs are available for eight machines, including the IBM and the new NEC."
Blank adds, "What we've done is design a machine independent language, which is emulated like Pascal on all the different machines, and optimized it for doing what we do. It's very compact. For example, on the mainframe, Zork I was about six hundred thousand bytes. That same game in its entirety is seventy-five thousand bytes on micros. So we get a factor of eight compression."
Berez then points out as we step aside for the bear, that "this gives us a tremendous marketing edge. If you can show us any machine that will sell a reasonable number of copies, we'll adapt our games to run on it."
The Infocom parser is also a unique and somewhat mystifying tool. Blank defines a parser as "a module of a program that takes English input and distills it into smaller pieces of information. The parser has three outputs. It understands a verb, a direct object, and an indirect object. If I say, 'Put the brown bag into that trophy case' " (pointing to the one behind us), "the parser looks at that and basically comes out with three things: put is the verb, brown bag is the direct object, and trophy case is the indirect object. Thus, where most adventures are limited to a parser that can only handle two words, our parser takes full sentences."
Old Pictures in New Frames. Blank also explains that Zork on the Apple was quite different in spots from the original mainframe version. Zork I is two-thirds from the original and one-third brand-new, Zork II is half and half, and the just-released Zork III is one-third old and two-thirds new. A black and white checkerboard figure looks over your shoulder as Blank discusses the depths of subtlety that Zork plumbed.
"If you're really in a place that's very foreign, very mysterious, nothing should be taken at face value. Everything should be explored," Blank advises. "For example, in Zork II, the Oddly-Angled Room was sort of a joke, like the way Beethoven would write a scherzo. The minuet movement was a very formal thing, the third movement in symphonies. Beethoven turned it into a scherzo, which means joke, and did funny things to a movement that is normally a very rigorously style thing.
"We decided, in the same way, that this maze was going to be different. You," Blank says, meaning us, "think you know how to do mazes. Here's a maze that you can't map. It's basically to tell people, 'Don't always think that a maze is a maze.' "
One of the big breakthroughs with Zork was the development of multisolutions to the same problem. The cyclops in Zork I could be put to sleep by feeding him the right food, or, if you ever happen to chance upon saying Ulysses, he would flee from the room. Blank offers a rare piece of Zork trivia. "Do you know that there is actually a clue to saying Ulysses or Odysseus? The commandment in the prayer book in the temple starts off with saying: 'Commandment 29,160--Oh ye who go about saying unto each other, "Hello Sailor." ' If you look at the first letter of each line going down the side, it reads 'Odysseus'!"
Beware the Fork. Hearing a mind-curdling scream from behind an unmarked door, you open the door onto a view of a demonic black temple. "What's that?" you yell. Blank swiftly closes the door, saying, "it is not time for you to know."
Somewhat shaken by this vision of blighted evil, we turn for comfort to a more familiar subject: Infocom's revolutionary breakthrough in the blossoming field of computer novels, Deadline.
"Deadline was our hardest project yet to implement," Blank comments. "People moving around--the idea that characters can walk a path by themselves. You can interrupt them and sometimes change their future moves completely, like switching train tracks. If you ask them about something, their answers depend on what you know. If you've shown them this, but not that, you'll get one answer. If you've shown them that, but not this, you'll get a different answer. And, if you've shown both things to the person, you'll get a third, entirely different answer. If you ask someone about something before it happens, they'll claim to have no knowledge of it.
"In Deadline, things are waiting for you to do something. You are an active force. It's possible that if you don't do anything in the game, nothing will really happen. There are thirty different endings possible. The newspaper headlines at the end convey how close to the best possible solution you were. When you finally satisfy this very finicky jury, you get, on the screen, the author's summary of what actually occurred.
"We pride ourselves on anticipating the things people will try to do, so that there'll be meaningful responses in the game and not just the old standard 'nothing happens.' Deadline is our best effort to date in that regard. It adds a sense that there is more going on in the game than what you see."
This Way to Debriefing. Lebling rejoins us, the natives having dropped him off from canoeing on the Charles River. He's excited about Infocom's first science fiction adventure, Starcross.
"Just as Mark is almost entirely responsible for Deadline, Starcross is mostly mine," Lebling smiles. "I have always been a science fiction fan and have wanted to do an adventure in the genre. That's one of the things I really like about Infocom. We figure out what we really want to do, rather than design games by market demand. I'm in this to have fun. It would be nice also for Infocom to make lots of money and be very successful, but I couldn't work if I wasn't having fun doing it. I love writing these games--much more than I enjoy playing them.
"Starcross was a real joy to write and should be a lot of fun for people to play. The puzzles are science fiction puzzles, not adventure puzzles. We did not want to do a 'Zork in Space' game. Starcross is intended as an entry level game for people who like science fiction but who haven't played many adventure games before."
The airlock puzzle that begins Starcross is worth the price of the whole game for its cleverness of design. Even the packaging of the game is extraordinary: the disk comes in a large flying saucer that's reusable for storage--and it really flies. Enclosed is a special color space map and a detailed manual. The map figures prominently in getting the game started.
Let 'Em Eat Disks. Referring to this and the excellent Deadline package, Berez puts forth the innovative idea that "merchandising can be a deterrent to piracy. People want all the goodies inside. That's the position Infocom is taking in the marketplace, in the sense that we're discouraging piracy by making the add-ons so attractive that people are willing to pay for them rather than settle for a pirated disk without the paraphernalia."
The outer door of the Starcross suddenly clangs shut and you hear the airlock recycling. Then the inner door opens in front of you and out steps the surprise you'd been told to expect: a cyborg. Michael Berlyn lights the room with his elfin grin. Now many of the strange things you've been seeing and experiencing throughout your visit begin to become crystal clear.
Sitting in the shade of the two ancient stone guardians of the Underground Empire, sharing a box lunch, Berlyn tells you of his journey to Infocom.
"It all started when Marc and I were introduced at Applefest. We hit it off so well that we decided right there to do a collaboration. Later on it looked like it would be better to go beyond collaboration and actually join up with him. I'm very glad to be here."
The future and its promises excite Berlyn. "I have two science fiction games planned. One is a collaboration with Dave and Marc. We intend to introduce some brand-new elements into the adventure genre with this game. I'm also working on a book called After the Change--six hundred pages long, so far.
"Being offered a position in this company is the pinnacle of my career. I'm into squeezing and wringing every ounce of energy, experience, creativity, and application out of myself that I can. I perceive Infocom as giving me the space, the time, and the opportunity to do just that."
Great excitement reigned at infocom upon Berlyn's decision to join them. This was not just a sudden, unplanned happening. Berez explains: "Since the summer of 1979, we've realized that our whole system lends itself to the talents of a professional writer. We've wanted to bring in people who could put most of their effort into writing good stories. Michael Berlyn--a published science fiction author with computer experience--is ideal. It's possible that what we're producing will become the novels of the future, and you're not going to want to read novels written by programmers. So we're providing new tools for writers to use to get their points across in new ways."
Blank discusses those tools and Berlyn: "Having Mike Berlyn here is one of the biggest pluses we have for the future. He's an excellent writer. He has lots of great ideas for sophisticated games that couldn't be implemented within the restrictions of Basic. With access to our technology, he'll be able to develop them. We're trying to extend the concept of the adventure story to where the player is a character and feels like a character, a part of the story, and not just like some being solving problems."
What Dreams May Come. Your lamp flickers; your remaining time here is short. If we're to learn about the future of Infocom, we must listen now. Berez complies.
"Al and I have just completed a business plan--about two inches thick. It establishes where we are today and where we want to go--the old five-year plan.
"In the future, we'll branch out from entertainment into other areas. For next year, there are five entertainment programs scheduled.
"One reason we're successful is the broad spectrum of computers on which our products run and the large installed user base that entails. Yet our transportability, for all its breadth, reaches only one-third to half of all the personal computers. The marketplace is growing and hungry, and our plans are to provide for it."
Your lamp is very dim and you can hear the baying of the grues in the distance. So, with a last whiff of the Perfect Rose, we bid Infocom adieu. Marc Blank leads us to a cell and waves as he sets a dial and pushes a button outside. As we pass through the Great Door, we hear the Wizard mutter an "F" word, and we leave with a warm glow in our hearts that will never leave.
Sleuthing for Spirits
The winding halls of Infocom seem to be peopled by the spirits of games. Not only to ghosts from Infocom's own games roam, but so do some from other games by Infocom's people.
How many allusions to these illusions can you spot?
Be the adventurer to identify the most references--obvious, subtle, obscure, even punned---and to place each in its real game, and you'll win the infocom game of your choice or $25 credit toward any Softalk advertiser's product.
In case of ties, precision will reign. If there are still ties, the random number generator will work overtime.
Send your answers to Softalk Ghosts, Box 60, North Hollywood, CA 91603, by November 15, 1982.
Thanks to Paul R. Santa-Maria for transcribing and donating this article.