Brass Lantern
the adventure game web site


David Lebling Interview

by Stephen Granade and Philip Jong

If you've played a few of the old Infocom adventures, chances are you've seen Dave Lebling's work. He was one of the implementors who created Dungeon, the mainframe adventure game which later became Zork I, II, and III. He wrote or co-wrote Enchanter, Starcross, Spellbreaker, Suspect, The Lurking Horror, and James Clavell's Shogun. Both he and fellow implementor Steve Meretzky were admitted to the Science Fiction Writers Association for their games. Since Infocom went away, though, he's not been doing games; instead, he's been doing more down-to-earth work with companies such as Avid.

You can imagine my excitement when the opportunity to ask him a few questions came my way. Philip Jong, proprietor of Adventure Collective, was in contact with Dave and asked me to do a co-interview.

Philip Jong [P.J.]: Your career at Infocom began during your days at MIT's Laboratory for Computer Science. What inspired you to become an interactive fiction author and game designer? How was it related to your research at MIT at the time?

Dave Lebling [D.L.]: The inspiration was the Crowther/Woods Colossal Cave Adventure, which hit MIT in late '76 or early '77. In addition, I had been playing D&D a lot, and was really enjoying it and thinking about creating D&D scenarios, so I was primed for interactive fiction. I had already written several games at MIT. I wrote a bunch of graphics games (vector graphics on a machine called an Imlac), including a port of the old Spacewar game; a graphical version of Hunt the Wumpus (some people think of Wumpus as the first "adventure" game); the original Maze game (with Greg Thompson), which was, as far as I know, the first FPS and the first networked FPS as well. I had also worked with Marc Blank and Tim Anderson on a Trivia game which had a database of thousands of questions submitted from all over the old Arpanet.

All of this stuff was only peripherally related to my research, which had to do with subjects such as Email, Morse Code decoding, Office Automation and programming tools. A lot of the game stuff was strictly after-hours fun, although it did help improve the language we used (Muddle, aka MDL), and was a test-bed for application packages that were used in the serious work. There was some interest in natural language parsing and understanding in our Morse Code project, and that may have had an effect on Zork as well.

P.J.: The original mainframe version of Zork was coauthored by Marc Blank, Bruce K. Daniels, Tim Anderson, and you. What were the key technological limitations in bringing your vision onto the computer screen back in 1977?

D.L.: The biggest limitation was that no one had tried to do an adventure game before in a systematic, relatively clean way in what we considered a reasonable language (all reasonable languages were Lisp or relatives of Lisp such as Muddle). We actually didn't see the source code for Adventure until later, but we knew it was written in Fortran, and the MIT attitude about Fortran was less than flattering! What we tried to do was write object-oriented code before that term was current, and to implement some idea of inheritance (so that objects could be specialized), but we did it without really knowing a lot about how those things (which were still in their infancy) should work. It came out kind of messy in the first go-around, as you might imagine, but we kept tweaking it (Tim Anderson was an indefatigable improver of the code). When Marc Blank and Joel Berez designed the Z-machine, they therefore had a lot of knowledge of how to do it more cleanly (from a language point of view) and came up with a really elegant, simple, machine.

As Zork grew bigger, the major impediment was that Muddle had a limited address space, laughable by today's standards, of about a megabyte (256K 36-bit words). Muddle was both the interpreter and the runtime environment for running the mainframe Zork, so this was a problem both in developing the game and in running it. The Muddle team had come up with some very ingenious ways of extending this using memory mapping, but it was a real limitation as the game grew bigger. It was a problem for debugging (which had to be done with the interpreted code, which was much larger than the compiled code), and for compiling (some of the modules were pretty large), and finally for creating and linking the "executable." We also grew dissatisfied with the speed we were getting, and spent a lot of time thinking of ways to speed up execution. Tim Anderson and Chris Reeve invented a faster calling sequence for Muddle functions that ended up helping a lot.

P.J.: There had been numerous speculations regarding the origin of the word "Zork." For the record, who among the "Infocom Imps" came up with this name? Where is the exact origin of the word "Zork"?

D.L.: I'm pretty sure it was Marc Blank who first applied the word to the game. The word itself was current as an exclamation or nonsense word (like "foo" and "bar") around the lab. Programs in the ITS operating system were had to have six-letter or fewer names, and it was pretty common to use a placeholder name when working on something new. I think Marc used "TS ZORK" as the placeholder, and it stuck.

I think "Frobozz" was similar, of a variant of "foobar." Bruce Daniels was, I think, largely responsible for its ubiquity in the early parts of Zork.

We briefly changed the name of the game to "Dungeon" (which was my bad idea, I sheepishly admit), then changed it back after TSR (the D&D people) threatened us with a lawsuit over it. MIT's lawyers squashed them like bugs but we decided we liked "Zork" better anyway. The widely distributed Fortran version of Zork was written during the period when the game was called Dungeon, which is why that version is often called Dungeon.

P.J.: When you and other members of the MIT's Dynamic Modelling Group founded Infocom in June of 1979 (two years after the debut of Zork), what was the driving idea behind the formation of the company? What were some of the growing pains Infocom had to go through during its initial years?

D.L.: Making giant seething gobs of money? Hmm, something more than that. There was interest in taking some of the expertise we had developed at the lab and making a company out of it. The funny thing was that initially no one was quite sure exactly what we would do! We kicked around a lot of ideas, most of them rather dull business-oriented things (a document-tracking system was one of them, for example) before considering a commercial version of Zork.

The major impediment was that no one was sure you could fit Zork on a microcomputer, and so Marc and Joel went off, thought about it, and then came back and convinced everyone else that it was possible. There was still some skepticism even then, and some resistance from the founders who were more interested in business products, but it worked. I believe the Z-language was the first wide use of disk paging on micros, certainly the first in games. It was what made it possible to fit the 84K bytes or so of Zork I into a much smaller real memory.

Infocom started pretty much as a self-funded, "garage-shop" (well, den- or kitchen table- or spare room-shop) operation. It grew slowly as people moved from MIT to it, and as we hired people to handle the business side of the operations. The corporate attitude, at least among the game writers and the whole engineering staff, was to work hard but have a lot of fun. Especially as the company grew, there were people much more concerned about being businesslike. This was, of course, before the advent of the web and the dot-coms, which to some extent legitimized the "geek" lifestyle and its disdain for the corporate ethos. There was an ungoing tension between the "businesslike" people and the "fun" people, even in the Cornerstone group as it came into existence and grew.

It's pretty much an article of faith among Infocom fans that Cornerstone was an evil thing that was forced on us and ultimately killed us. Actually, we were good friends with the two guys who mostly invented Cornerstone (Brian Berkowitz and Richard Ilson), they were colleagues from the lab, and we were excited about applying the same technology that made Zork so portable (a byte-code compiler, interpreter and runtime) to business products. There was also an understanding that the prices and margins on business products were much higher than on games. Not everyone agreed with all that, but it was not a case of management doing something that everyone under them opposed.

What actually did kill us was not Cornerstone per se, but bad decisions about how to fund, package and market Cornerstone. Essentially we ended up using the income from the games to fund the business product, to the detriment of both. One way of looking at it was that the venture capital market was skittish about investing in a company that did both games and business products, but the management team was scared that if they split the company or spun off a Cornerstone, Inc. that they would lose control to the venture capitalists.

P.J.: Among the many interactive fiction titles you authored or coauthored, which single title do you feel best represents the pinnacle of your contribution to the area of interactive fiction? Why?

D.L.: There are things I love about all the games I wrote. Picking one is like picking which child you like best; both impossible and impolitic. "Zork II," "Enchanter," and "Spellbreaker" are my favorites, if forced to choose. I think "Spellbreaker" has some of my best prose, and I'm very proud of the magic system I designed for "Enchanter." "Zork II" is a favorite because of the Wizard of Frobozz, one of my favorite characters. I also like the thinly-disguised MIT I created for The Lurking Horror. As you can see, I have a hard time choosing.

"Spellbreaker" has the nastiest puzzles, bar none, so if I have to pick one I'll go with it. I think "Zork 2" or "Enchanter" for sheer playability.

P.J.: Among the many interactive fiction titles you authored or coauthored, which single title do you feel best represents the pinnacle of your contribution to the area of interactive fiction? Why? Which one of your works was the biggest disappointment to you? Why?

That would have to be "Shogun," which was written as Infocom was dying around me. The project was originally Activision's idea (they were very big on licensed titles), and they sold it as a collaboration like the one we had done with Douglas Adams on "Hitchhiker's Guide."

In practice, James Clavell was interested in computers but far too busy to actually make it a collaboration. I had been excited about working with him (and in fact we did spend some time kicking ideas around), but the book is just too huge and too full of carefully-orchestrated plot to adapt into an adventure game.

What I ended up writing is a sort of "Scenes From..." game. It has some good parts, I think (a couple of timing puzzles that are very intricate), and Donald Langosy did a superb job on the graphic add-ins. However, as a whole I have to consider it a failure. The technology wasn't (and still isn't) ready for a plot-and-character-heavy adventure game.

Stephen Granade [S.G.]: If you had the chance to redo any of your Infocom games, which one would you change? What would you do differently, or would you avoid the game entirely?

D.L.: I'd redo the infamous Baseball Diamond puzzle in Zork II, which has been an object of universal hatred ever since its implementation. I'd have loved to have done "The Lurking Horror" as a larger-size game (it was almost the last of the "small" games which had to fit in 84k bytes of disk space). Some good scary stuff got cut out of it or never implemented due to the size restrictions.

I'd also love to redo "Suspect," using a more modern language that supports object classes better than the Z-language did. The things I was trying to do there (have each character be sensitive to the state of the mystery) would be much easier to do, and the result would be much more realistic.

None of the games was perfect; it would be fun to clean up any of them.

P.J.: What was your opinion of Infocom's merger with Activision? Did your opinion change when Jim Levy stepped down as CEO of Activision shortly after the merger and Bruce Davis took over? How did the working environment at Infocom change after the merger with Activision?

D.L.: What is a drowning man's opinion of a life-preserver? We were essentially out of money when we merged, so Activision saved us from bankruptcy. They seemed like a bunch of people who knew what they were doing in the videogame market and wanted to join with people who knew what they were doing in the PC game market, so it seemed like the merger had a reasonable chance of success. In practice, it worked less well, as I'm sure all mergers do. We lost independence, and they didn't understand Infocom's methods of doing business. We always thought of ourselves as more like a book publisher than a movie production company. That is, we cultivated a backlist, we produced a limited number of titles, we cared deeply about quality. Activision was more like a movie company. They got a game up on the screen and two weeks later got another one up on the screen. They were impatient with the idea of slower but steady long-term sales (and to be fair, the industry was changing at this time in a way that made the old Infocom model work less well). However, they didn't manage the transition from our way of doing business to theirs very smoothly. They began to have financial problems not long after the spate of acquisitions which included Infocom.

When Bruce Davis took over Activision, things went downhill even faster. He had opposed the merger in the first place, and eventually demanded money from the Infocom officers and founders to try to recover some of the purchase price, using what we thought were patently ridiculous arguments. This led to lawsuits in both directions. I think it's fair to say that he began the fracas partly as a way of deflecting attention from his own awful management.

The suit had, as you might imagine, a negative effect on morale. In addition, as Activision sank they cut budgets and jobs at Infocom even more extensively than at their home office. They pressured us to do movie and book licenses, and to do more titles (which of course flooded the market). They had made some very ill-advised acquisitions, and none of them were working out, and they were running out of cash. Ultimately Joel Berez left (Marc Blank and Al Vezza were long since gone) and was replaced by Joe Ybarra, who they appointed as President (or was it general manager?). Joe had little management experience, which didn't help matters.

S.G.: What's your opinion of Activision's continuation of the Zork series? Zork Grand Inquisitor in particular referenced many things from the original series--any qualms about seeing your old creation in the hands of strangers?

D.L.: I thought they were reasonable attempts at graphical adventures, but in general the areas in which they were strong were the areas least like the Zork series, and vice versa. They were exploiting a valuable property they owned, and I know that the authors were trying to do something in keeping with what we had done, but the results were disappointing.

P.J.: Among all people you have collaborated with, who gained your highest respect? Why?

D.L.: I mostly collaborated with people inside Infocom. I always found Marc Blank to have a game-sense or puzzle-sense most like mine, and we collaborated easily on several games, most notably Enchanter. I think I was most in awe of Steve Meretzky, whose reservoir of ideas never seems to run dry, and who has always been a meticulous and dedicated craftsman about his work, and cares deeply about every aspect of it. All this while denying that he is a programmer.

P.J.: What did you do after you left Infocom? A while back we heard you were with Avid (a company that does special effects). Do you still work for them?

D.L.: Actually Avid pioneered digital film and video editing, and is still the leader in that industry. They also do special effects, as you mention. I did a lot of different things at Avid, including a visual language for representing stories, a television newsroom news broadcast management system, and some work on a consumer video editing product.

However, I left Avid in November of 1999 and joined a startup called Ucentric Systems, which is building a Home Server platform. This is a device that combines a gateway, router, web-server, telephone system, video services and a bunch of applications in one inexpensive box.

P.J.: Do you believe the principles of design should be the same for graphic adventure titles as for text adventures? Why or why not?

D.L.: No. A visual medium has very different story-telling principles than a textual one. I don't think that there has yet been a good graphical adventure. Most of the possible candidates have been illustrated text adventures. The closest thing may have been Myst and its sequels, but as adventures they are very primitive, no matter how spectacular their graphics. I usually point to the distinction between books and movies when asked this question. Stephen King's "The Shining" is an excellent book, and Stanley Kubrick's "The Shining" is also an excellent movie, but the things that make them excellent are very different.

S.G.: Back in 1996 you made a few brief posts to the two interactive fiction Usenet newsgroups. Have you ever considered becoming a more active member of that community?

D.L.: Every now and then I read the groups, but frankly, I don't have that much spare time and keeping up with any newsgroup is hard.

P.J.: The appeal of the adventure genre (including interactive fiction) has dramatically lessened since the golden days of Infocom. What do you think are the driving forces behind the apparent demise of the genre? Do you believe the success of Myst played any role in the slump in the adventure genre?

D.L..: It's pretty clear to me that sophisticated graphics and sound are pushing games more and more in the direction of Hollywood, and away from the printed book. A text adventure is an interactive book, and as such you can write puzzles that take advantage of the 100,000-plus word vocabulary of English. In "old school" adventure games you could perform hundreds of different actions, and the result was (sometimes) some pretty clever puzzles and immersive stories. Unfortunately, even the most advanced graphics games have a much smaller vocabulary. (In Myst, the vocabulary consisted primarily of "walk" and "press button," for example.) On the other hand, Myst's incredible-for-their-time graphics attracted a lot of casual players who then couldn't solve the puzzles. I don't think very many of them bought any other adventure games.

P.J.: Do you believe the adventure genre is dead or dying?

D.L.: As a commercial proposition, yes. It is caught between the incredible advancements that have been made in graphics, and which continue to improve almost daily, and the stagnation of the advancements in actually communicating with your computer. Just to give an example, Everquest, which I am completely addicted to, is crammed with puzzles in the form of "quests." But the interaction with the quest NPCs is at a level even below that of Colossal Cave. The "parser" just looks for slightly varied forms of the output text from the NPCs, text that they actually enclose in brackets to make it completely obvious what to say in reply.

One rarely sees an adventure of any sort on the best-seller lists. I suspect that many adventure-game players have moved to the MMORPGs like Everquest, which provide some of the rich environment of adventure games and make a stab at the puzzle-solving aspect of them as well. They provide some realism in the character interaction by having most of the other characters be real people. Others have moved to the world-building games like Civilization or more recently, Black and White, or out of computer gaming altogether.

Infocom was very lucky in that it caught the technology and demographic curves perfectly in its early days. I think that both curves have moved on, and they don't intersect any longer in a place that makes adventure games commercially viable.

S.G.: Do you play any of the newer text adventures?

D.L.: Sadly, no. I've played a little of some of the shareware ones I saw discussed on the newsgroups, but as you embarrassingly point out, that was a few years ago. Most of my limited gaming time these days is devoted to Everquest.

P.J.: What is your current gig?

D.L.: I work in Java, Javascript, JSP and HTML designing and coding applications for the Home Server we are creating at Ucentric Systems. This is a Linux-based universe, which is fun, and Java is very Muddle or Z-language-like with its Collection classes and byte-code compiler, although it looks like C or C++.

P.J. and S.G.: Where do you see yourself five years from now? What about writing a new piece of interactive fiction yourself?

D.L.: Retired on a beach in the Caribbean, sipping pina coladas? Probably not. I suspect I'll be doing what I'm doing now, which is writing code for something that interests me. That's how I got started on Zork, and that's pretty much what I've always done. I've really been lucky in that regard; I haven't had to spent very much time in my career working I things I wasn't interested in, or was bored by.

As for writing interactive fiction, I suspect it's unlikely. I would actually be tempted if I got a chance to design stuff for an MMORPG, especially Everquest (for reasons of familiarity). It would be a challenge to come up with things to make those games more adventure-like, to enhance their puzzle components.

I think the space of what can be done in text adventures has been well-explored by a variety of very creative people (by no means all of whom worked at Infocom). It would take, I fear, a qualitative leap in the development language or environment to expand that space. We never got very good at doing conversation, for example. There's a long way to go before realistic conversations exist in games. We were okay but not spectacular at giving people more than one way to solve a problem. You need a more advanced input method to solve that one: People are just not that interested in typing to the game to simulate physical actions. A virtual-reality suit would solve that but they're a long way off.

No one has yet solved the primary problem of adventure games, which is, what happens when the player doesn't do what you expected? Once progress is made on that one, it might be fun to write an adventure game again.

About Us | Contact Us | Technical Info | History
Copyright © 1997-2010, Stephen Granade.