The Eastgate School of “Serious” Hypertext

A quarter of a century after Ted Nelson first coined the term, hypertext finally stepped into the spotlight in 1987. As we’ve seen in an earlier article, the primary driver of its long-delayed public recognition was Apple’s HyperCard, which in the wake of its premier at the MacWorld show in August went on to become the product of the year in the eyes of most industry pundits. But concurrent with the HyperCard hype were a number of other, smaller developments — enough to convince one that hypertext’s newfound fame might be down to more than just the whim of a major corporation, that it might be in some more organic sense an idea whose time had simply come.

The American Association of Computing Machinery, the oldest and arguably the most respected learned society devoted to computing in the world, had decided to hold a conference on hypertext well before HyperCard was more than a Silicon Valley rumor. When the conference actually took place in November of 1987, however, it could only benefit from the HyperCard excitement, which gave it a sense of relevance that stuffy academic conferences all too often don’t manage to capture. While many things were discussed over the course of those few days, the conference would go down in history for the debut of Storyspace, the first tool explicitly designed for authoring hypertext narratives on a personal computer, and for that of afternoon, a story, the first work ever to label itself a “hypertext novel.” These twin debuts also mark the beginning of what would become known as the Eastgate school of self-described “serious” hypertext, one of the less accessible — in both the figurative and, today, the literal sense — movements in the history of digital narratives.

The co-creator of Storyspace and the author of afternoon was Michael Joyce, a professor of language and literature at Jackson Community College in Jackson, Michigan. Shortly after completing his first print novel in 1981, Joyce had bought his first computer, an Apple II, and immediately been captivated by what he saw as a whole new world of writing possibility. Responding to what he described as his frustration with the limitations of linear storytelling, he cultivated an eclectic web of friendships to pursue his interest in exploring new narrative structures enabled by computers. Most prominent among this group were Howard Becker, a sociologist at Northwestern University and fellow Apple enthusiast (he provided Joyce with a steady flow of pirated games, including many Infocom titles); Natalie Dehn, a researcher at Yale’s Artificial Intelligence Lab; and Jay David Bolter, a classicist at the University of North Carolina who was investigating generative storytelling on computers as a sideline. After upgrading to the new Apple Macintosh soon after its release, Joyce and Bolter, the latter of whom was a self-taught programmer, began working in earnest on Storyspace. From the beginning, Bolter took advantage of the Macintosh GUI to make the system accessible to non-programmers like Joyce. A snapshot of the work in progress from 1985:

The program currently represents structure as a map or network of rectangular cells and straight lines. Cells are units of text that may range in size from one character to 30,000. The author creates cells, labels them, positions them on the screen using the mouse, and attaches text. Stacking cells inside other cells indicates hierarchical relationships, while drawing and labeling lines from one to another indicates associative links. The author may then use the created structure to control or review the presentation of the text.

The pace of Storyspace development accelerated that same year when Joyce and Bolter won a grant from the Markle Foundation, allowing them to employ other programmers on the project.

The current Storyspace version 3 is not at all different conceptually from the system that Joyce and Bolter came up with in the 1980s. Nodes of text are represented as onscreen cells, to be connected together by the author using the mouse. The system is very accessible for non-programmers, but, because there is no real facility for tracking state, much less for modeling a world “behind” the surface text, it can also be very limiting. How much the character of the works created with Storyspace was down to ideology in the abstract and how much was down to ideology molding itself to the limitations of the technology is a little questionable.

In years to come, Michael Joyce and his fellow proponents of serious hypertext would seem willfully determined to disassociate themselves from existing commercial software — and especially from computer games. It’s interesting therefore to note that Joyce and Bolter’s original descriptions of Storyspace to potential funders didn’t describe it strictly as the tool for the academic avant garde that it would eventually become. The Markle Foundation funded Storyspace’s development based on a pitch that emphasized its applicability to business and to more mainstream, Choose Your Own Adventure-style branching narratives.

Still, someone had to be the first to make something in Storyspace. A new development system of any stripe can always benefit from a killer app to demonstrate its capabilities, and Storyspace was no exception, as Joyce’s friend Howard Becker pointed out to him: “One thing that will really help nerds like me see how to use this will be a couple of good examples, spelled out in real detail, and included on the disk. Like a story by you…” afternoon, a story, the most famous, most read, and most analyzed work of the Eastgate school, was thus created with the rather practical goal of simply showing off what Storyspace could do to potential investors and customers. Joyce began to write it in March of 1987, and had completed it in time to bring it to Chapel Hill, North Carolina, in November for the ACM hypertext conference, hosted by Jay David Bolter’s employer the University of North Carolina. Joyce and Bolter gave out copies of afternoon to conference attendees after their presentation on Storyspace.1 History hasn’t recorded in any detail the computer scientists’ reaction to Joyce’s daunting work of postmodern literature, but we’re on firmer ground on the subject of the authoring system he had used to create it: Storyspace became the hit of the conference.

Following that rapturous reception, Joyce and Bolter, joined now by a University of North Carolina computer scientist named John B. Smith, formed a company they named Riverrun and started pitching Storyspace to software publishers. They talked for some time with Brøderbund, who went so far as to lend them hardware and expertise to further the project after the Markle Foundation’s funding ran out. But in February of 1989 Brøderbund bowed out in response to, as historian Matthew Kirschenbaum puts it, “lingering confusion over exactly what the tool did and who its potential audience was.” Brøderbund was seemingly skeptical whether this group of academics was truly capable of creating a product that would appeal to the mainstream of Middle America, the commercial sweet spot Brøderbund was almost uniquely adept among their peers at reaching with products like The Print Shop and the Carmen Sandiego series.

Having been rejected by Brøderbund and the rest of the consumer-software industry, who were going through troubled times and growing ever more risk-averse thanks to the Nintendo onslaught, Joyce, Bolter, and Smith turned to Mark Bernstein, founder of a tiny company called Eastgate Systems dedicated to researching future applications of hypertext. In 1990, Eastgate published Storyspace at last, and also published afternoon, a story as their debut work of hypertext fiction. Both garnered modest interest in the mainstream press as curiosities and possible harbingers of the future. Indeed, Eastgate had big plans for the future. Occupying some hazy middle ground between software publisher and book publisher, they would for much of the 1990s publish multiple works of hypertext fiction and nonfiction each year alongside the Storyspace software that was used to create them. For a time they even published their own magazine, The Eastgate Quarterly Review of Hypertext, full of theory on the one hand and news about their latest products on the other, like a highbrow version of Infocom’s old New Zork Times. In 1992, Robert Coover, like Michael Joyce a print novelist turned hypertext evangelist, published an opinion piece in The New York Times that served as a manifesto of sorts for what the Eastgate school of writers believed would become a major — indeed, potentially revolutionary — literary movement. We’ll return to that a little later. But first, more on afternoon, a story itself and the works it spawned.

Authors of the Eastgate school, who were almost universally academics working in the humanities, saw themselves as pushing the literary novel to the next stage of its formal evolution. For this reason, many or most of their works have as much or more to do with explicating certain ways of thinking about literature and literary criticism as they do with plot, character, or any of the other attributes of traditional novels. Their works are steeped in the post-structuralist school of literary criticism, which is itself an outgrowth of the postmodern philosophy of thinkers like Jacques Derrida. While I do understand that an extended discussion of such topics probably won’t set your hearts aflutter with anticipation, I do think a little background is necessary to an understanding of what the Eastgate school was all about. So, please bear with me while I set the stage as painlessly as I can manage.

The road to the Eastgate school really does begin with Jacques Derrida, who from the 1960s until his death in 2004 remained the preeminent voice of postmodernism as it applied to literary criticism and many other endeavors. It was Derrida who invented the concept of deconstruction — a concept that, like many of the concepts associated with postmodern philosophy, seems virtually impossible to fully define. In writings that must stand as some of the most impenetrable ever committed to paper, Derrida himself did such a baffling, self-contradictory job of it that he’s often been accused of not knowing quite what it was himself, and of attempting to obscure that lack through sheer weight of verbiage. The problem was only complicated in later days by the mainstream media’s latching onto the term and using it constantly as little more than a synonym for “analyze.”

Still, if we stay safely at the shallow end of the pool, deconstruction can be a very straightforward, noncontroversial idea: the idea that one can learn much about a text by teasing out the unexamined assumptions of its author. Herman Wouk, to take an example, doubtless considered himself quite an enlightened man when he published The Caine Mutiny in 1951, but his condescending descriptions of the titular minesweeper’s black kitchen staff says much about the racist attitudes of his time. At this level, then, deconstruction implies little more than a skeptical reading between the author’s lines, and a willingness to seek context outside of the work itself.

At the deeper end of the pool, however, we come to the claims that nothing can ultimately mean anything at all. Derrida rejected the notion, underpinning in different ways both religion and science, that there is an absolute Truth out there somewhere which we can approach if not reach via earnest inquiry. In Derrida’s view, any absolute Truth must belong to the realm of metaphysics — a realm in which he refused to believe. Instead of Truth, he saw a multiplicity of individual, subjective “truths,” hierarchies of constructed meaning — thus meanings ripe for deconstruction — tied to hierarchies of social power. Deconstructionism has always walked in hand in hand with Marxism and other radical political ideologies. Just as Marxism envisions an end to the privileges of authority, deconstructionist thought seeks an end to the privilege accorded to the author as the final authority on her work’s meaning.

At the risk of being accused of getting too cute or playing games of gotcha!, I can’t resist pointing out the logical contradiction inherent in the supposed objective Truth that there is no objective Truth. More seriously, though, the idea that all meaning is constructed and subjective is one that will doubtless continue to strike each incoming group of humanities undergraduates as a profound revelation, and to strike those of us who have been around it for a few years — those of us who haven’t become Derrida scholars, that is — as the most tedious of hobby horses to continue flogging. It seems to me that the problem with radical deconstructionism and, indeed, postmodernism in general is that it’s very hard to know what to really do with them. What’s the point of saying anything if you don’t believe it’s possible to say anything that bears any relationship to any Truth outside itself? Much rationalization has been done in an attempt to avoid the nihilism to which postmodernism would seem inevitably to lead, but the arguments have never struck me as terribly convincing.

Of course, any attempt to fully capture Truth in writing, whether on the grand scale of history or the empathetic scale of an individual character, must inevitably fail at some level, must run afoul of subjectivity and the limits of the author’s cognition and experience. This is a Truth that any competent writer or historian — and I do like to believe I manage to be both on a good day — must always keep in mind. Still, the point of the endeavor is the striving, the point is to come as close as you can to the ideal of a captured Truth. If you don’t believe there is anything out there to be striven for, why bother? The debate is not strictly an academic one. Taken to an extreme, a refusal to believe in the existence of verifiable facts is not just absurd but actively dangerous, as the current president of my country is so masterfully demonstrating as I write these words.

But now let’s turn our attention back to afternoon, a story, a work steeped in postmodern thought, to see what effect those patterns had on this work of literary hypertext.

I try to recall winter. < As if it were yesterday? > she says, but I do not signify one way or another.

By five the sun sets and the afternoon melt freezes again across the blacktop into crystal octopi and palms of ice  — rivers and continents beset by fear, and we walk out to the car, the snow moaning beneath our boots and the oaks exploding in series along the fenceline on the horizon, the shrapnel settling like relics, the echoing thundering off the far ice.

<Poetry > she says, without emotion one way or another.

Do you want to hear about it?

And so, with this passage that could only have issued from an overwrought teenager or a tenured academic, we begin our “story” — a word, one has to assume, that Joyce means ironically, for replying that yes, we do want to hear about it, yields anything but a straightforward story. This is not an exercise in “What do you want to do next?” It’s rather a web of allusions and musings, with no foregrounded action at all. As near as I’ve been able to divine through much feverish clicking, our story, such as it is, is that of Peter, a man who has just witnessed a car crash that may have killed his ex-wife and his son. If we are persistent enough, we may eventually arrive at a node that seems to say that Peter himself may be responsible for their deaths in some way. But that’s the closest we can ever get to any sort of resolution; this alleged “story” of more than 500 nodes is not only nonlinear but endless, every node always looping back onto other nodes.

“A thin young man with a lavender penis and huge, swollen balls,” huh? Don’t threaten me with a good time!

In terms of interface, afternoon must strike us today as a strange beast, and it’s a little hard to determine how much of that strangeness is down to conscious intent and how much is down to the era when it was first created, well before our modern expectations for a hypertext interface had been set in stone. Those sections of the text which lead to other sections — in Joyce’s preferred parlance, those “words that yield” — are not highlighted in any way. Indeed, in later editions of afternoon every word in the text will lead you somewhere else, albeit all too often to one of the same set of uninteresting cul de sacs. There is at least a back button for when you get caught in one of these, along with a forward button that will yield a default next node if you don’t wish to choose one — but, typically for this work that seems so self-consciously designed to stymie and frustrate its reader, traversing the entire text using the default option only winds you up in one of those uninteresting cul de sacs. And then there are also “yes” and “no” buttons, which you can use to respond to occasional explicit questions or just click for the hell of it at any time to go somewhere else. Given that afternoon consists of more than 500 nodes, and that the relationships between them all are intuitive at best, random at worst — certainly anything but logical — trying to get a sense of it all is a fairly monumental endeavor.

But people have certainly tried, and therein lies a noteworthy tension between what afternoon believes itself to be and what it actually is. Michael Joyce and many others of the Eastgate school were always determined to disassociate their work from games, even “literary” games like those of Infocom.2 Yet to say that afternoon isn’t a game in the same sense that Infocom’s interactive fictions are games goes against the way that almost everyone responds to it. (Whether adventure “games” and similar interactive works really are games in the same sense that competitive zero-sum exercises are games is a separate discussion which we’ll have to leave unaddressed today.) Confronted with this word salad, we want to figure it out, to make sense of it, to find out what the hell Peter is on about. Thus there is a puzzle to be unraveled here, an implicit ludic challenge to be confronted. Even much of the academic writing on afternoon obsesses over this process of figuring it out. In this sense, then, afternoon is not so far from an Infocom game — or for that matter a mystery novel — as Joyce might prefer to believe. Whether it constitutes a good game or puzzle is of course another question entirely.

And on that subject, I have to be honest: afternoon, a story bores the living daylights out of me. If it was interested in empathetically exploring the feelings of Peter, it might have been a moving work. If it was interested in letting the reader get to the definitive bottom of what really happened to Peter’s ex-wife and son, it might have been an intriguing one. If it could have shown even a flash from time to time of self-awareness or humor instead of remaining so relentlessly, pretentiously po-faced, it might at least have been a little more likeable. But, sadly in my view, Michael Joyce isn’t interested in doing any of these things. What Joyce is interested in is, in the words of critic Janet Murray, “intentionally ‘problematizing’ our expectations of storytelling, challenging us to construct our own text from the fragments he has provided.” Yet the text he has provided is so leaden and dull that the only type of interest such an exercise can muster is theoretical interest in the mind of a post-structuralist literary critic. I find this sort of self-reflexive art — art about nothing more than the process of its own creation, or the process of its own reception — to be a betrayal of art’s potential to move and change us. “Problematizing our expectations of storytelling” is a thin foundation on which to build a work of deathless literature.

So, we return now to that zeitgeist-capturing New York Times piece of 1992, written by Robert Coover, who would go on to become something of an elder statesman for the Eastgate school. It was entitled “The End of Books.” Like so many zeitgeist-capturing pieces, it comes across as almost hilariously dated today, but nevertheless remains the logical next stop for anyone trying to understand what the Eastgate school was all about. First, the article’s opening. Afterward, in the spirit of turnabout being fair play, we shall indulge in a little of what some might refer to as deconstruction of our own.

In the real world nowadays, that is to say, in the world of video transmissions, cellular phones, fax machines, computer networks, and in particular out in the humming digitalized precincts of avant-garde computer hackers, cyberpunks and hyperspace freaks, you will often hear it said that the print medium is a doomed and outdated technology, a mere curiosity of bygone days destined soon to be consigned forever to those dusty unattended museums we now call libraries. Indeed, the very proliferation of books and other print-based media, so prevalent in this forest-harvesting, paper-wasting age, is held to be a sign of its feverish moribundity, the last futile gasp of a once vital form before it finally passes away forever, dead as God.

Which would mean of course that the novel, too, as we know it, has to come to its end. Not that those announcing its demise are grieving. For all its passing charm, the traditional novel, which took center stage at the same time that industrial mercantile democracies arose — and which Hegel called “the epic of the middle-class world” — is perceived by its would-be executioners as the virulent carrier of the patriarchal, colonial, canonical, proprietary, hierarchical and authoritarian values of a past that is no longer with us.

Much of the novel’s alleged power is embedded in the line, that compulsory author-directed movement from the beginning of a sentence to its period, from the top of the page to the bottom, from the first page to the last. Of course, through print’s long history, there have been countless strategies to counter the line’s power, from marginalia and footnotes to the creative innovations of novelists like Laurence Sterne, James Joyce, Raymond Queneau, Julio Cortazar, Italo Calvino and Milorad Pavic, not to exclude the form’s father, Cervantes himself. But true freedom from the tyranny of the line is perceived as only really possible now at last with the advent of hypertext, written and read on the computer, where the line in fact does not exist unless one invents and implants it in the text.

“Hypertext” is not a system but a generic term, coined a quarter of a century ago by a computer populist named Ted Nelson to describe the writing done in the nonlinear or nonsequential space made possible by the computer. Moreover, unlike print text, hypertext provides multiple paths between text segments, now often called “lexias” in a borrowing from the pre-hypertextual but prescient Roland Barthes. With its webs of linked lexias, its networks of alternate routes (as opposed to print’s fixed unidirectional page-turning) hypertext presents a radically divergent technology, interactive and polyvocal, favoring a plurality of discourses over definitive utterance and freeing the reader from domination by the author. Hypertext reader and writer are said to become co-learners or co-writers, as it were, fellow travelers in the mapping and remapping of textual (and visual, kinetic and aural) components, not all of which are provided by what used to be called the author.

Coover comes across like a caricature of the pretentious academic with his name-dropping irrelevant asides to demonstrate his erudition (“what Hegel called ‘the epic of the middle-class world'”) and his painful attempts to capture the spirit of William Gibson (“the humming digitalized precincts of avant-garde computer hackers”). The language of deconstruction and post-structuralism is everywhere. The conventional reader is forever being “dominated” and “tyrannized” by the “patriarchal,” “colonial,” “canonical,” “proprietary,” “hierarchical,” “authoritarian” author. But, even setting the loaded language aside, is she really so much under the author’s thumb? It seems to me that the reader always has the ultimate veto power over the author in the form of the ability to put the book down and not read it anymore. If she choose to continue, to follow the line to the end, she is doing so willingly, and can hardly be considered a subjugated victim.

Further, is the “tyranny of the line” truly a tyranny? Far from being compelled to follow the author “from the first page to the last,” the reader is always free to skip around in a text, to read the parts that interest her in the order she prefers. Why else do tables of contents and indexes exist? Those of us who made it through a graduate program in the humanities were all forced by the sheer amount of reading that is assigned to learn how to delicately but mercilessly fillet a book in order to extract exactly what we need from it in the minimum length of time. Far from being subjected to the author’s tyranny, we learned to rip the beating heart out of the author’s little darling and dissect it on our desktops. It’s a skill that still serves me well in the work I do today, as the bloody remains of this article’s list of sources lying strewn around me right now will attest.

And then is the best way to achieve a “plurality of discourses” really to try to stuff them all into a single nonlinear book? Wouldn’t it perhaps make more sense to just, you know, read several books?

But most problematic of all is the article’s core premise/prophecy: that hypertext will lead to the “end of books.” The end of print, Coover tells us, will “of course mean that the novel, too, as we know it, has to come to its end.” In reality, there’s no “of course” about this proposition at all. Coover demonstrates a colossal failure of imagination in conflating the literary form of the novel with print, the physical medium on which novels were almost exclusively delivered in 1992. As the thriving modern ebook market demonstrates, the medium is not necessarily the message in this case. Let’s give him the benefit of the doubt, though, and look past even this confusion. As long as we’re talking about tyranny, we should note that Coover’s essay never asks what real-world readers actually want. Do they really wish to become “co-learners” and “co-writers”? I think most would reply that writing is what they’re paying the author to do, and that the author had damn well better learn her subject on her own time, before she starts trying to sell them a book about it.

Which isn’t to say that there isn’t room for interactive texts, and now more so than ever. As many readers of this blog are doubtless aware, choice-based stories like those published by Choice Of Games have achieved a certain degree of commercial success as the heirs to the Choose Your Own Adventure gamebooks of yore, while more poetic hypertext explorations created with tools like Twine reach their own audience of readers. Granted, the audiences for both styles of work remain by all indications fairly small, but they nevertheless dwarf the numbers that have ever read the literary hypertexts of the Eastgate school. So, while we are far, far from any end of books — and thank God for that! — there obviously is a readership for hypertext fiction, one that may even grow considerably in the future. The fact that the works which people are reading are so divorced from the works of the Eastgate school has everything to do with the failure of said school to write hypertext novels that people want to read. Summing up the heyday of the Eastgate school, scholar Adam Hammond says that “the most enduring legacy of this period is theoretical rather than creative.” It’s not really meant as a criticism of the works themselves, yet it’s hard for me to think of a better one.

Writing about the status of serious hypertext as of 2016, Hammond notes that it “is one of the few digital literary forms that can be plausibly regarded as dead”: “The works that seemed so revolutionary and world-altering were not only seldom read in their own time, but are today literally unreadable, languishing in antiquated software and hardware formats inaccessible to the contemporary reader.” Eastgate soldiers on somehow, still selling Storyspace — in fact, a new version was recently released — alongside a collection of hypertext novels that virtually all date from those halcyon days of the 1990s. Yet it’s hard to figure out to whom they could be selling this stuff as of 2017. They’ve done, as Hammond says, an horrendous job of keeping up to date with the changing times. I recently bought three hypertext novels from Eastgate: afternoon, a story, King of Space, and Victory Garden. The first of this group is the most famous work ever published by Eastgate, while the last is among the two or three next most famous. And yet none of them will Just Work on a modern computer. afternoon, a story comes the closest: it comes on a USB stick that will work on Mac OS X — but only on versions released prior to 2011. Victory Garden, meanwhile, won’t run on a 64-bit operating system of any stripe, while King of Space requires, incredibly, a version of Mac OS prior to OS X. (“1 M of memory and hard disk required,” announces the packaging.) I’m a fairly technical guy, and thus can get all this stuff running, but how many other prospective readers can make the same claim? When I placed my order, I was greeted with emails explaining the situation and asking if I really wanted to pay all this money for these relics, which Eastgate still sells for $25 or more apiece. There then followed more elementary questions about delivery methods and the like. Everyone was very nice, but I couldn’t help myself from conjuring a scene that ought to be a Saturday Night Live skit if it isn’t one already, taking place in a disused office space somewhere inhabited by a couple of dozing employees: “Hey, wake up! We actually have a customer!”

As should be obvious by now, I have little truck with most of what I’ve seen of the Eastgate school. I’m always skeptical of writers who make a point of telling me they are writing “serious” works, for it always leads me to think about the greatest writer I’ve ever read, William Shakespeare, who never laid claim to being more than a working playwright feeding the public’s appetite for new entertainments. Writers in my opinion should simply write, and leave the judgments to the critics and, most of all, to history. Given this bias of mine as well as all of the other problematic (to me) rhetoric surrounding the Eastgate school, and given that it so conspicuously failed to set the world on fire as promised in Coover’s article, you might well be asking why I’ve bothered to write all these words about its origins. It’s a fair question, and one to which I will, if you’ll bear with me just a little longer, give three answers.

The first answer is that there is some real historical interest here in the context of more contemporary choice-based narratives. Storyspace was the first of its kind, the urtext of systems like Twine and the Choice Of Games engine. While it has generally been used to create works of a very different character than those more recent systems, there is no reason why it couldn’t create, say, a contemporary take on Choose Your Own Adventure full of concrete choices — “I want to go here and do that!” — instead of more abstracted explorations of a writerly space.

The second answer is that the movement simply deserves an evaluation — or an obituary if you must — from someone outside its orbit. I may very well not be the person who is best-qualified to provide that evaluation, but for right now anyway I’m all it’s got.

But it’s the third answer that may be the real answer. This blog has always been, among other things, a chronicle of my personal journey through the history of interactive entertainment. And the fact is that, while I’ve seen the big flagship works and not been overly impressed, much of what languishes in the Eastgate catalog is completely unknown to me, and much of it sounds rather intriguing despite that whisper from my worse if wiser self that tells me I’m probably in for more disappointment. I wonder for instance about the aforementioned King of Space, a hypertext novel I’ve purchased but haven’t yet tried; in addition to what certainly reads like an extant and even intriguing plot, it contains, Eastgate tells us sotto voce, “elements of gaming.” (Shhh!) So, I do plan to dig into this and other Eastgate hypertext novels, to find out whether there are hidden away in their catalog works with some real meat on their bones and life flowing through their veins. I’ll continue to explore, and if I find that to be the case I’ll report back here. If not, this may well be my one and only article on the subject — mission accomplished, duty to history satisfied. We shall see.

In the meanwhile, how about we talk about an adventure game next time? Sound good to you? Well, at this point it sounds pretty good to me too.

(Sources: Literature in the Digital Age: An Introduction by Adam Hammond; Mechanisms: New Media and the Forensic Imagination by Matthew G. Kirschenbaum; “Postmodernism and Science Fiction” by Andrew M. Butler, from The Cambridge Companion to Science Fiction; Narrative as Virtual Reality and Avatars of Story by Marie-Laure Ryan; Cybertext: Perspectives on Ergodic Literature by Espen J. Aarseth; Hamlet on the Holodeck: The Future of Narrative in Cyberspace by Janet H. Murray; The New York Times of June 21, 1992.)

  1. afternoon, a story would go through multiple revisions over a period of years, but these would largely be to polish the existing text and, most of all, to add many more links between the extant textual nodes. In terms of word count, in other words, the version of afternoon that Joyce passed out at the conference was substantially the same as the one that is available today. 

  2. The most amusing of the rare collisions between literary hypertext and the hobbyist interactive-fiction community must be the history of the Usenet newsgroup Being apparently unaware that for many thousands of people the term “interactive fiction” meant text adventures — especially the text adventures of Infocom — literary-hypertext practitioners tried to appropriate the term for themselves around the time of Eastgate’s release of afternoon, setting up to discuss their own very different interests. It was promptly invaded by vastly greater numbers of text-adventure fans wanting to talk about the old Infocom games, whereupon the literary-hypertext people departed in a huff, leaving the newsgroup, alongside its companion group, to become the central discussion forum of the Interactive Fiction Renaissance of the 1990s. 


Posted by on February 10, 2017 in Digital Antiquaria, Interactive Fiction


Tags: ,


For text-adventure fans confronting the emerging reality of a post-Infocom world, AGT was a godsend, allowing amateurs for the first time to create games that at a quick glance might appear to match those of Infocom. Still, AGT was far from the answer to every prayer, for the fact remained that it would have to be a very quick glance indeed. David Malmberg, the developer of AGT, was ingenious and motivated, but he was also a self-taught programmer with little background in the complexities of programming languages and compiler design. He had built AGT by adding to Mark Welch’s profoundly limited GAGS system for creating generic database-driven text adventures a scripting language that ignored pretty much all of the precepts of good language design. What he had ended up with was almost a sort of technological folk art, clever and creative and practical in its way, but rather horrifying to anyone with a deeper grounding in computing theory. The best argument in favor of AGT was that it worked — basically. While the system was undoubtedly more capable than anything that had been available to hobbyists before, it still didn’t take much poking at an AGT game before the rickety edifice’s seams began to show.  A number of authors would push the system to unforeseen heights, in the process creating a number of classic games, but it was obvious that AGT could never work quite well enough to create games as polished as those of Infocom. To do that, a language would be needed that was truly designed rather than improvised. Enter Michael J. Roberts with his Text Adventure Development System, or TADS.

Roberts had first started programming on his school’s DEC PDP-11 system in the 1970s at the age of 12, and thereafter had ridden the first wave of the PC revolution. In those early days, text adventures were among the most popular games there were. Like so many of his peers, Roberts studied the BASIC listings published in books and magazines carefully, and soon started trying to write adventures of his own. And like so many of those among his peers who became really serious about the business, he soon realized that each game he wrote was similar enough to each other game that it made little sense to continually reinvent the wheel by writing every one from scratch. Roberts:

It occurred to me that lots of the code could be generalized to any adventure game. So, I tried to write a little library of common functions — the functions operated on a set of data files that specified the vocabulary words, room descriptions, and so on.

This was a nice approach in some ways; the idea was that the game would be described entirely with these data files. The problem that I kept running into was that I’d have to write special-purpose functions for certain rooms, or certain commands — you couldn’t write an entire game with just the data files, because you always had to customize the library functions for each game. What I really wanted was a way to put programming commands into the data files themselves, so I didn’t have to modify the library routines.

Once you start putting procedural code into data files, you essentially have a programming language. At first, I tried to avoid the work of writing a real language interpreter by making the language very limited and easy to parse. That was better than just the data files, but it was tedious to write programs in a limited language. I eventually saw that you really needed a good language that was easy to use to be able to write decent games.

Roberts, in other words, was discovering for himself the limitations and inelegancies that were inherent to a system like AGT — the limitations and inelegancies of grafting a scripting language onto a generic database engine.

But it wasn’t until he went off to the California Institute of Technology that his experiments progressed further. Despite his official status as a physics major, CalTech offered plenty of opportunity for a motivated young hacker like him to immerse himself in the latest thinking about programming languages and compiler design. In the air all around him was computer science’s hot new buzzword of “object-oriented” design. By allowing the programmer to gather together data and the code that manipulates that data into semi-autonomous “objects,” an object-oriented programming language was an ideal fit for the problems of constructing a text adventure. (Indeed, it was such an ideal fit that Infocom had developed a heavily object-oriented language of their own in the form of their in-house adventure-programming language ZIL years before the buzzword came to prominence.) Following yet another trend, Roberts based his new adventure language’s syntax largely on that of C, the language that was quickly become the lingua franca of the world of computer programming in general.

On the theory that a worked example is worth a thousand abstractions, and following the precedent I set with my article on AGT, I’d like to show you a little bit of TADS code taken from a real game. You may want to refer back to my AGT article to contemplate the comparisons I’m about to make between the two languages. Taken together, the two articles will hopefully lead to a fuller understanding of just how TADS evolved the text-adventure-programming state of the art over AGT.

The game we’ll be looking at this time is Ditch Day Drifter, a perfectly playable standalone adventure that was also used by Mike Roberts as his example game for TADS learners. Once again following my AGT precedent, I’ll focus on that most essential piece of equipment for any old-school adventurer: a light source. Here we see Ditch Day Drifter‘s flashlight.

flashlight: container, lightsource
    sdesc = "flashlight"
    noun = 'flashlight' 'light'
    adjective = 'flash'
    location = security
    ioPutIn( actor, dobj ) =
        if ( dobj <> battery )
            "You can't put "; dobj.thedesc; " into the flashlight. ";
        else pass ioPutIn;
    Grab( obj ) =
         *   Grab( obj ) is invoked whenever an object 'obj' that was
         *   previously located within me is removed.  If the battery is
         *   removed, the flashlight turns off.
        if ( obj = battery ) self.islit := nil;
    ldesc =
        if ( battery.location = self )
            if ( self.islit )
                "The flashlight (which contains a battery) is turned on
                and is providing a warm, reassuring beam of light. ";
                "The flashlight (which contains a battery) is currently off. ";
            "The flashlight is off. It seems to be missing a battery. ";
    verDoTurnon( actor ) =
        if ( self.islit ) "It's already on! ";
    doTurnon( actor ) =
        if ( battery.location = self )
            "The flashlight is now on. ";
            self.islit := true;
        else "The flashlight won't turn on without a battery. ";
    verDoTurnoff( actor ) =
        if ( not self.islit ) "It's not on. ";
    doTurnoff( actor ) =
        "Okay, the flashlight is now turned off. ";
        self.islit := nil;

Unlike the case of our AGT example, for which we had to pull together several snippets taken from entirely separate files, we have here everything the game needs to know about the flashlight, all in one place thanks to TADS’s object-oriented design. Let’s step through it bit by bit.

The first line tells us that the flashlight is an object which inherits many details from two generic classes of objects included in the standard TADS library: it’s both a container, meaning we can put things in it and remove them, and a light source. The rest of the new object’s definition fleshes out and sometimes overrides the very basic implementations of these two things provided by the TADS library. The few lines after the first will look very familiar to veterans of my AGT article. So, the flashlight has a short description, to be used in inventory listings and so forth, of simply “flashlight.” The parser recognizes it as “flashlight” or “light” or “flash light,” and at the beginning of the game it’s in the room called “security.”

After this point, though, we begin to see the differences between TADS’s object-oriented approach and that of AGT. Remember that adding customized behaviors to AGT’s objects could be accomplished only by checking the player’s typed commands one by one against a long series of conditions. The scripts to do so were entirely divorced from the objects they manipulated, a state of affairs which could only become more and more confusing for the author as a game grew. Like those created using many other self-consciously beginner-friendly programming languages, AGT programs become more and more of a tangle as their authors’ ambitions grow, until one reaches a point where working with the allegedly easy language becomes far more difficult than working with the allegedly difficult one. Contrast this with TADS’s cleaner approach, which, like Infocom’s ZIL, places all the code and data pertaining to the flashlight together in one tidy package.

Continuing to read through the TADS snippet above, we override the generic container’s handling of the player attempting to put something into it, specifying that this particular container can only contain one particular object: the battery. Then we specify that if the player removes the battery from the flashlight when the flashlight is turned on, its status changes to not lit — i.e., it goes out.

Next we have the flashlight’s “long description,” meaning what will happen in response to the player attempting to “examine” it. TADS allows us to insert code here to describe the flashlight differently depending on whether it’s on or off, and, if it’s in the latter state, depending on whether it contains the battery.

Finally, we override the generic light source’s handling of the player turning it on or off, to tie the written description of these actions to the specific case of a flashlight and to reckon with the presence or absence of the battery. Again, note how much cleaner this is than the AGT implementation of the same concepts. In AGT, we were forced to rely on several different versions of the flashlight object, which we swapped in and out of play in response to changes in the conceptual flashlight. In TADS, concept and code can remain one, and the steps necessary to implement even a huge adventure game can continue to be tackled in relative isolation from one another.

Instructive though it is to compare the divergent approaches of the two systems, it is important to state that TADS wasn’t created in reaction to AGT. Computing communities were much more segregated in those days than they are today, and thus Roberts wasn’t even aware of AGT’s existence when he began developing TADS. Beginning as a language tailored strictly to his own needs as a would-be text-adventure author, it only gradually over the course of the latter 1980s morphed in his mind into something that might be suitable for public consumption. What it morphed into was, nevertheless, something truly remarkable: the first publicly available system that in the hands of a sufficiently motivated author really could create text adventures as sophisticated as those of Infocom. If anything, TADS had the edge on ZIL: its syntax was cleaner, its world model more thorough and consistent, and it ran in a virtual machine of its own that would prove as portable as the Z-Machine but was free of the latter’s brutal size constraints.

As TADS was rounding into this very impressive state, Roberts set up a company with a friend of his named Steve McAdams. In tribute to Roberts’s degree in physics, they called it High Energy Software, and, following in the footsteps of David Malmberg’s little AGT enterprise, planned to sell TADS as shareware through it. Version 1.0 of TADS was released in September of 1990, alongside two games to show it off. One was the afore-referenced freebie example game Ditch Day Drifter, while the other was Deep Space Drifter, a bigger game released as a shareware product in its own right. Both games tread well-worn pathways in terms of subject matter, the former being yet another “life at my university” scenario, the latter a science-fiction scenario with some of the feel of Infocom’s Starcross. Both games are a little sparse and workmanlike in their writing and construction, and some elements of them, like the 160-room maze in Deep Space Drifter, are hopelessly old school. (It’s not a maze in the conventional drop-and-map sense, and the puzzle behind it is actually very clever, but still… 160 rooms, Mike? Was that really necessary?) On the positive side, however, both games are quite unusual for their era in being scrupulously fair — as long, that is, as you don’t consider the very idea of a huge maze you have to map out for yourself to be a crime against humanity.

But undoubtedly the most ambitious and, in their way, the most impressive of the early TADS games came not from High Energy Software but rather from a pair of University of Maryland students named David Leary and David Baggett, who started a company they called Adventions to sell TADS text adventures via the shareware model. Of all the folks dabbling in shareware text adventures during the early 1990s, it was Adventions who made the most concerted and sustained effort at building a real business out of it. Their flagship series came to encompass three big unabashed Zork homages — Unnkulian Underworld: The Unknown Unventure, Unnkulian Unventure II: The Secret of Acme, and Unnkulia Zero: The Search for Amanda — alongside Unnkulia One-Half: The Salesman Triumphant, a free snack-sized sampler game.

When the first Unnkulia game was released remarkably quickly on the heels of TADS itself — Mike Roberts can’t recall for sure, but believes Leary and Baggett likely developed it with an early beta copy of the system — it stood as easily the most immediately impressive amateur text adventure ever. The text was polished in a way that few text-adventure developers outside of Infocom, whether amateur or professional, had ever bothered to do, being free of the self-conscious meta-textual asides and atrocious grammar that had always marked the genre. Adventions’s text, by contrast, looked like it had actually been proof-read, and possibly several times at that. Likewise, the game took full advantage of the sophisticated TADS world model to offer puzzles of an intricacy that just wasn’t possible with a tool like AGT. The first Unnkulia game and those which followed were almost in a league of their own for some time in all these respects.

On the other hand, though, the Unnkulia games strike me as curiously unlikable. You can get a good idea of their personality just by looking at their names. If the name Unnkulia — be sure to say it out loud — strikes you as hilarious, congratulations, you may have found your new favorite series. If it instead just strikes you as stupid, as it does me, perhaps not so much. (I admit that my attitude may be affected by having to type the damn thing over and over again; no matter how hard I try, I just can’t seem to remember how to spell it.) Much of the humor inside the games for some reason involves “cheez” — and yes, it’s spelled just like that. The humor has always been, at best, polarizing, and I have no doubt on which side I stand. In addition to just not being all that funny, there’s a certain self-satisfied smugness about the whole enterprise that rubs me the wrong way. At the risk of over-personalizing my reaction to it, I’ll say that it feels like the work of two young men who are nowhere near as witty as they think they are. In short, there’s something about these games that I find insufferable.

In terms of design, the Unnkulia games are an equally odd brew. It’s clear that they’ve been quite rigorously tested — another thing that sets them apart from most text adventures of their era — and they’re free of mazes, guess-the-verb puzzles, and the other most-loathed aspects of their genre. Yet the puzzle design still isn’t all that satisfying. There’s an obsession with hiding objects behind, under, and inside unlikely things — an obsession which is ironically enabled by the TADS world model, which was the first to really allow this sort of thing. Sometimes, including in the very first room of the very first game, you even have to look twice to find everything. Hiding surprises in plain view is okay once or twice, but Baggett and Leary lean on it so hard that it quickly starts to feel lazy. When they do get more ambitious with their puzzles, however, they have a tendency to get too ambitious, losing sight in that peculiar way so many text-adventure authors have of how things actually work in a real physical environment. Let me offer a quick example.

So, let’s set up the situation (spoilers ahoy!). You have a bronze plate, but you need a bronze coin to feed to a vending machine. During your wanderings in search of that among other things, you come upon the room below. (These passages should also convey some more of the, shall we say, unique flavor of the writing and humor.)

Inner Temple of Duhdism

This chamber is the temple of Duhdism, the religion of the ancients. It's rather a letdown, after all Kuulest told you about it. A small altar with a round hole in the center is in the center of the chamber. Carved in stone on the far wall is some writing, the legend of Duhdism. The only exit from this chamber is back to the east. You feel at peace in this room, as if you could sleep here - or maybe you're just kind of bored.

>read writing
"The Legend of Duhdha and the Shot to Heaven:

One fine summer day, Duhdha was loading a catapult with rocks. When his students asked what he was doing, the great Duhdha just smiled and said, "Something real neat." Soon, the catapult was full, and Duhdha pulled the lever as his students looked on. The stones crushed the annoying students, leaving the great man to ponder the nature of mankind. Not only did the rocks eliminate distraction from Duhdha's life, but they fell to the ground in a pattern which has since become a standard opening for the intellectual game of "Went." Since then, the altar at the Temple of Duhdha fires a small stone into the air soon after a worshipper enters, to honor Duhdha - who taught his students not to ask stupid questions and to pretty much just leave him alone."

>x altar
The altar is about two feet by one foot, and about three feet tall. There's a small round hole in the exact center of the top surface. The altar is covered with rock dust. There's nothing else on the Duhdist altar.

Time passes...
A rock shoots into the air from the hole in the altar, shattering on the ceiling and spreading rock dust on the altar.  From the outer chamber, you hear the old monk cry "Praise Duhdha!."

We obviously need to do something with this rock-spewing altar, but it’s far from clear what that might be, and fiddling with it in various ways offers no other clues. Putting things on it has no effect on either the thing that’s just been put there or the rocks that keep flying out — except in the case of one thing: the bronze plate we’re carrying around with us.

>put plate on altar

Time passes...
A rock shoots from the altar at high velocities, puncturing the plate in the center. The rock shatters on the ceiling, spraying rock dust. You hear a tinkling sound as a tiny bronze disc falls on the floor. From the outer chamber, you hear the monk shout "Praise Duhdha!"

*** Your score has just changed. ***

When you reach this solution either by turning to the walkthrough or through sheer dogged persistence — I maintain that no one would ever anticipate this result — you might then begin to wonder what physical laws govern the world of Unnkulia. In the world we live in, there’s no way that the flying rock would punch a perfect hole neatly through the middle of a bronze plate that happened to just be lying on the altar. Without something to hold it in place, the plate would, of course, simply be thrown into the air to come down elsewhere, still intact. The ironic thing is that this puzzle could so easily have been fixed, could have even become a pretty good one, with the addition of a set of grooves or notches of some sort on the altar to hold the plate in place. Somehow this seems emblematic of the Unnkulia series as a whole.

Like everyone else who dreamed of making a living from shareware text adventures, Leary and Baggett found it to be a dispiriting exercise, although one certainly can’t fault them for a lack of persistence. With some bitterness — “It’s disappointing that although there are so many IF enthusiasts out there, so few are willing to pay a fair price for such strong work,” said Baggett — they finally gave up in time to release their final game, 1994’s very interesting The Legend Lives! — more on that in a future article —  for free before exiting from the text-adventure scene entirely.

The same dispiriting lack of paying customers largely applied to makers of text-adventure languages as well. Mike Roberts estimates that his rate of TADS registrations peaked “on that same order” as David Malmberg’s pace of 100 AGT registrations per year, or “maybe a little lower. I’d remember if it had been much higher because I’d have been spending all day stuffing envelopes.” Whatever their respective rates of registration, far more AGT than TADS games continued to be released in the early 1990s. While that may have struck some — not least among them Mike Roberts — as disappointing, the reasons behind it aren’t hard to divine. AGT had a head start of several years, it had an annual competition to serve as an incentive for people to finish their games, and, perhaps most of all, it presented a much less daunting prospect for the non-programming beginner. Its kind and gentle manual was superb, and it was possible to get started making simple games using it without doing any programming at all, just by filling in fields representing rooms and objects. TADS, by contrast, offered a manual that was complete but much drier, and was on the whole a much more programmer-oriented experience. The initial learning curve undoubtedly put many people off who, had they persisted, would have found TADS a much better tool for realizing their dreams of adventures in text than AGT.

Some time after the Adventions guys and David Malmberg gave up on their respective shareware products, Roberts and his partner Steve McAdams also decided that they just weren’t making enough money from TADS to bother continuing to sell it. And so they made TADS as well free. And with those decisions, the brief-lived era of shareware interactive fiction passed into history.

But, despite the disappointments, Mike Roberts and TADS weren’t going anywhere. Unlike Leary, Baggett, and Malmberg, he stayed on the scene after giving up on the shareware thing, continuing to support TADS as open-source software. It took its place alongside a newer — and also free — language called Inform as one of the two great technical catalyzers of what some came to call, perhaps a little preciously, the Interactive Fiction Renaissance of the mid- to late-1990s. So, I’ll have much, much more to write about TADS and the games made using it in years to come. One might even say that the system wouldn’t really come into its own as a creative force until Roberts made that important decision to make it free.

Still, the importance of TADS’s arrival in September of 1990 shouldn’t be neglected. Somewhat underutilized though it may initially have been, it nevertheless remained the first development system that was capable of matching Infocom’s best games. If the amateur scene still largely failed to meet that lofty mark, they could no longer blame their technology. On a more personal note, the emergence of Mike Roberts on the scene marks the arrival of the first of the stalwarts who would go on to build the modern interactive-fiction community. We’re still in an era that will strike even most of the most dedicated fans of modern interactive fiction as murky prehistory, but some people and artifacts we recognize are beginning to emerge from the primordial ooze. More than a quarter of a century later, Mike Roberts and TADS are still with us.

(Sources: SynTax issues 17 and 36; SPAG issues 5 and 33; Mike Roberts’s interview for Jason Scott’s Get Lamp documentary, which Jason was kind of enough to share with me in its entirety. And my thanks to Mike Roberts himself, who answered many questions personally via email.

Ditch Day Drifter and Deep Space Drifter are available on the IF Archive for play via a TADS interpreter. The Adventions games are available there as well.)


Tags: , ,

The Spellcasting Series (or, How Much Ernie Eaglebeak is Too Much Ernie Eaglebeak?)

In an earlier article, I described Steve Meretzky during his peak years at Infocom as “second to no one on the planet in his ability to craft entertaining and fair puzzles, to weave them together into a seamless whole, and to describe it all concisely and understandably.” But his talents encompassed more than just good puzzle design. As an Infocom Imp, he had the unique gift of taking potentially hackneyed or just plain dumb premises and making them subtly, subversively smart. Who would have expected Leather Goddesses of Phobos, his ribald low-brow sex romp, to prove so clever and joyous and downright life-affirming? Who would have expected Stationfall, on its face the most rote sequel Infocom ever made — “Bring back Floyd!” had shouted the masses, and Meretzky had obliged — to turn gradually and without any warning whatsoever into the creepiest, most oppressive game in their canon?

It’s because I know what he’s capable of at his best that I’ve always found Meretzky’s post-Infocom career a little disappointing. He did make good games after Infocom, but his track record from those years was far more mixed. During the 1990s, his dumb premises too often lacked the requisite touch of smarts to leaven the brew, and some of his design genius seemed to desert him. And sadly, his disappointments often smacked not of aiming too high, as he arguably had with his flawed would-be Infocom masterpiece A Mind Forever Voyaging, but rather from aiming too low, from being content to wallow in his established persona as a maker of wacky adventure games without adding that subversive spice that had elevated games like Leather Goddesses of Phobos and Stationfall to more interesting heights. Seen from this standpoint, his very first game after Infocom’s shuttering, Spellcasting 101: Sorcerers Get All the Girls, while a perfectly serviceable wacky adventure game in its own right, does rather portend the less inspiring phase of his career that was now beginning.

That said, when removed from the context of the games Meretzky was still to make, and particularly from that of the direct sequels to this game that were to follow, Spellcasting 101 is hardly a crime against adventure gaming. In fact, it was a wise if safe choice of a game to mark Legend Entertainment’s debut as a company and Meretzky’s as a freelance game designer. I already laid out the reasons for that in my last article: Steve Meretzky was well-known for his wacky comedy adventures; his Leather Goddesses of Phobos had been Infocom’s last big hit; Sex Sells in the abstract. Ergo Spellcasting 101. The first release of this very fragile new company was not the time to take too many artistic chances.

Spellcasting 101, then, finds Steve Meretzky sitting very comfortably in his wheelhouse, mixing unabashedly dumb humor — I might use the adjective “sophomoric,” but perhaps I should reserve that for the sequel Spellcasting 201 — with smart puzzles. You play Ernie Eaglebeak, the latest in adventure gaming’s long line of loser protagonists. He’s in love with a hot little number named Lola Tigerbelly, and dreams of going off to Sorcerer University to study magic, but, in a sex-reversed version of the Cinderella story, is prevented from doing either of these things (sorry!) by Joey Rottenwood, his evil stepfather. (With a name like that, I think old Joey should have been fronting a punk-rock band, but what do I know?) In the prologue, you in the role of Ernie defy your stepfather’s wishes in the matter of your education at least by sneaking off to Sorcerer University after your acceptance letter arrives in the mail.

The game’s a little more risque than Leather Goddesses of Phobos, not least in that it now has pictures, but it’s still the sort of thing that only the most hormone-addled teenager is likely to find genuinely titillating; the sex is still very much played for laughs, not for, shall we say, self-gratification. Bob Bates of Legend had a rule of thumb for deciding what was and wasn’t acceptable to portray in the pictures. It’s best described as the “Playboy cover test”: anything that might appear on the cover — not the inside pages — of a Playboy magazine was okay. Like Leather Goddesses and its own two sequels, Spellcasting 101 does let you play in “nice” or “naughty” mode. Indeed, some of the most amusing gags in the games come if you choose to become one of the approximately one percent of players who don’t immediately switch to “naughty” mode. Suddenly nights of passion are replaced by, say, a night spent playing gin rummy. (Any resemblance to my marriage after eight and a half blissful years is, I’m sure, purely coincidental.)

One could wish that Meretzky had been able to let the player play as a female, as he had in Leather Goddesses of Phobos, but the need to depict so much of what happens visually, along with a more complicated plot and a more fleshed-out protagonist, precluded that possibility. The enforced male gaze does open the games up to a charge of sexism that it’s pointless to even try to refute. (Just look at those box covers and screenshots!) I suppose one could try to argue that the Spellcasting series is really making fun of the males doing all the leering, in the style of the Leisure Suit Larry series, but the humor doesn’t even have as much edge as those games, making it hard to regard it as satire. In the Spellcasting games’ defense, however, it’s all so over the top that it’s also hard for me at least to take any of it seriously enough to work up much outrage. Certainly Legend got little to no backlash from offended consumers, leaving Steve Meretzky’s wish to author a truly controversial game, which had burned since his stridently anti-Reagan effort A Mind Forever Voyaging back in 1985, still unrealized.

The explicit political statement that opened Leather Goddesses of Phobos felt bracing in its time, but the constant back-patting and referencing of notable contemporary culture warriors in the Spellcasting games gets to be a bit much. We get it, Steve, you’re a real rebel against the Establishment with your wacky adventure games.

The humor in general is a little hit and miss. Where the first Spellcasting game in particular really shines, however, is the puzzle design — which, it must be said, is sometimes inseparable from the humor, and therein lies much of its brilliance. Many of the puzzles rely on wordplay, such as the unexpected use of a spell to “increase bust size.” In fact, there’s an entire extended sequence, “The Island of Lost Soles,” built around a premise that would have suited Nord and Bert Couldn’t Make Head or Tail of It perfectly. The inhabitants of the titular island have been entrapped within the various objects there. So, for example, “Blaize” is now trapped inside a campfire on the beach, and can be released only by casting a spell to “restore lost souls” on his true name. The usual problem with puzzles like this — and certainly the big problem with the should-have-been-a-classic Nord and Bert — is that they rely so much on the player’s native vocabulary and intuition. There are going to be some connections that even the most gifted player in both categories will simply never make, leaving her wandering hopelessly stuck. Spellcasting 101 solves that problem with a stroke of genius that’s disarmingly simple. Once you’ve wandered for a given amount of time without making further progress, a “hint fairy” will show up, saying something like “Have you seen Blaize around here?” With an actual name now to hand, you can search the island again, looking for the right object to which to apply it. Thereby does the game preserve the joy of making spontaneous intuitive connections without ever allowing you to get hopelessly hung up on those connections you can’t make. This, my friends, is what good design looks like.

Indeed, there’s a lot of attention paid in Spellcasting 101, as there would be in most of Legend’s later games to an even greater extent, to treating the player fairly. Even the combinatorial-explosion problem that made so frustrating Zork Zero, Meretzky’s final epic for Infocom, is addressed here by dividing the game into manageable chapters, each with its own discrete set of goals. Granted, some of the chapters are time-based, requiring you to plan your efforts around a constantly ticking clock. Solving these parts requires a fair amount of trial and error — i.e., figuring out what needs to be done over the course of several passes through a given chapter, then making a plan and bringing it all together via a final speed run. You will, in other words, be saving and restoring quite a lot despite the absence of more egregious design sins. Still, it should also be noted that Spellcasting 101‘s two sequels are much bigger sinners in this department. For long stretches of Spellcasting 101, the ticking clock disappears — something that most definitely can’t be said about the sequels.

So, this first Legend adventure game ever evinces a clear desire to be welcoming and accessible, if necessary at the expense of the sort of really intricate puzzles beloved by some of Infocom’s more hardcore disciples. Suffice to say that it’s a long, long way from the enormous, complicated, bafflingly non-linear Zork Zero. Kudos to Steve Meretzky, who was known to complain during Infocom’s latter days that their games were getting too easy, for recognizing the role his latest game needed to play in getting Legend off the ground, and for adjusting his design to suit the circumstances. Making due accommodation for the fact that this style of humor will never be to everyone’s taste, if there’s a holistic complaint to be leveled against the design it’s probably that the whole thing is just a little schizophrenic. What seems like it’s going to be a game about life at Sorcerer University suddenly transforms just as you’re settling into it into a series of vignettes, played on islands scattered all over the Fizzbuttle Ocean, that could have come from almost any adventure game. The Island of Lost Soles is a prime case in point: delightful as I find it on its own merits, it has nothing really to do with any other part of the game that houses it beyond yielding when all is said and done the requisite piece of the magical whatsit you’re trying to collect. The other island vignettes are less extended but equally isolated from one another and from the overarching plot of the game as a whole. Luckily, they’re all entertaining enough that the objection becomes more philosophical than impedimentary.

Call me overly sensitive if you must, but this scene from Spellcasting 201 is one of those that creeps me out just a bit. It also illustrates one of the inconsistencies in the series’s writing. Ernie Eaglebeak is the classic put-upon nerdy loser who couldn’t buy a date — until it’s time to score, whereupon all the sexy ladies suddenly have the hots for him.

Spellcasting 201: The Sorcerer’s Appliance, Meretzky’s 1991 sequel, does fix this problem; the entire game now takes place inside the confines of a greatly expanded Sorcerer University. (The extended map is explained as the result of “an extensive campus renovation and expansion program.”) Now in his sophomore year, Ernie will need to save the world entirely from within the confines of the university and its immediate environs this time, in between attending classes and pledging to the fraternity Hu Delta Phart — whose name, along with those of its companion fraternities Tappa Kegga Bru, I Phelta Thi, and Gramma Eta Pi, gives a pretty good idea of the sort of humor you’ll find in these games, for anyone still in doubt.

Unfortunately, the previous game’s focus on playability and accessibly falls by the wayside. In Spellcasting 201, which is divided into chapters each representing one day, you’re the constant slave of a ticking clock that runs in increments of no less than five full minutes per turn. (Ernie, it appears, may be not just slothful but a genuine sloth.) While none of the puzzle solutions are out-of-left-field howlers, they do often require lots of trial and error. And thanks to that ticking clock, trial and error in this context means save and restore, again and again and again. As your collection of save files mounts, you also have to contend with a strict inventory limit that forces you to juggle objects in just the right way as you go about each day.

Some of the puzzles seem almost consciously engineered to be as annoying as possible, as if Meretzky has forgotten the lesson, long since taken to heart by his old peers at Infocom, that annoying the player in the name of comedy is seldom a good design strategy. Perhaps the worst offender of all is a magical musical instrument called a moodhorn, whose proper operation also serves as part of the game’s copy protection. You can play moodhorn compositions that are found in a music book that accompanies the game to affect the mood of the people around you — songs like “Happiness Interlude,” “Fear,” “Drippingtreesap’s Love Theme No. 15,” and “Winter Cold.” It sounds fun enough in theory, with, in the best tradition of the Enchanter series that did so much to inspire the Spellcasting games, lots of opportunities for Easter eggs even where puzzles solutions aren’t forthcoming. In practice, however, it’s just excruciating. Each song requires entering a series of six commands for manipulating the moodhorn, typing things like “vomp plunger,” “trib high glupp key,” and “oscilloop half burm lever” one after another in sequence. Because the moodhorn seems like it ought to be useful in many situations, you’ll likely spend much of the game wandering around trying out the twelve separate compositions listed in the music book, looking for the one place where the moodhorn actually is useful. Have I mentioned how excruciating this is? Just to add the final dollop of absurdity to the exercise, the ticking clock means that every moodhorn composition requires fully thirty precious minutes to play. (Even Yes albums don’t have songs that long!)

Legend stepped up to VGA graphics with Spellcasting 301, but the girls continue to look like they were assembled from mismatched spare parts left over after making other girls.

The series wobbled to its anticlimactic close with 1992’s Spellcasting 301: Spring Break, in which Ernie and his fraternity buddies head for the beach. The ticking clock remains and is as annoying as ever, although this time you are granted the mercy of being able to leave some tasks incomplete and still finish the game, albeit at a cost to your final score. Otherwise, the pleasures and pains are largely the same as those of the second game. The one immediately obvious difference is that the pictures, including all those babes that never seem to be put together quite the way real women are, are now in 256-color VGA rather than 16-color EGA. Given how much of a piece the Spellcasting games — particularly the last two — really are, I’d like to use the remainder of this article to try to understand what went so wrong with the series as a whole rather than to dwell on the details of Spellcasting 301‘s individual failings.

If we’re looking for someone to blame for the Spellcasting series’s problems — besides poor Steve Meretzky, that is — our first  candidate has to be Ernie Eaglebeak, the hapless loser you’re forced to play in all three games. Simply put, he’s a horrid little twerp, the sort of kid that you want to drag down to the beach just so you can personally kick some sand in his face. Worse, he’s such a thoroughly uninteresting horrid little twerp. Even as odious adventure-game losers go, he’s no Leisure Suit Larry. And as Ernie Eaglebeak goes, so goes the rest of the cast of bimbos, dirty old men, and fraternity bro-dudes. The most sympathetic character in the series, a dottering old professor by the name of Otto Tickingclock, gets cuckolded and then accidentally killed by Ernie, who cares not a whit. There’s precious little warmth or joy to be found herein. It’s baffling to think that these games were written by the man who gave us Floyd, the first character in a text adventure that we ever really cared about. The death knell for any piece of fiction, interactive or otherwise, comes with the opposite reaction, when the reader says, “I really don’t care what happens to these people.” It’s hard to imagine anyone invested enough in Ernie Eaglebeak to give a damn about him.

As fiction, then, the Spellcasting series has its problems. Ditto as comedy. Various Implementors who worked at Infocom, as well as many commentators over the decades since, have described the strict limitations imposed by Infocom’s original 128 K Z-Machine as a counter-intuitive benefit to their games in that it forced authors to hone their works down to polished jewels, excising all of the weaker bits, retaining only the very strongest material. I’ve always been and to some extent remain a little skeptical of this thesis; as I’ve stated in previous articles, I think that some of the latter-day Infocom games in particular were trying to do a little too much in too small a space, and could have really used a few more kilobytes at their disposal. That said, the Spellcasting series makes a very strong counterargument for the value of restraint, stringently enforced if need be. With a system to hand for the first time that supported effectively unlimited amounts of text, Meretzky felt free to write… and write… and write. Some of the gags go on forever, to little if any comedic payoff. At Infocom, all this material would have been ruthlessly pared down by Meretzky’s fellow Imps and the testers, leaving only the genuinely funny bits behind. In this new order, however, it all just lies there on the screen, limp and lifeless as the comatose girl Ernie crawls into bed with in one of the series’s squickier episodes.

Cringe-worthy meta-textual comments like this one from Spellcasting 201 were par for the course in amateur AGT text adventures of the time, but they’re a little disconcerting to see in a $40 boxed commercial game. This sort of lazy writing would never have made it out of the first round of testing at Infocom.

Having broached the subject of Infocom’s working methods versus those Meretzky would utilize as an independent designer, I’d like to continue to follow that line of thought, for I think it might just lead us to the core reason that the Spellcasting series and so much of Meretzky’s other post-Infocom work pales in comparison to his earlier classics. Even after he became Infocom’s most famous Imp, Meretzky remained just one member of a creative collective which had the willingness and ultimately the authority to restrain his more questionable design impulses. This was the well-honed Infocom process for making great adventure games in action. I think that Meretzky needed that process — or at least a process involving lots of checks and balances — in order to turn out his best work. Legend did try as much as possible to duplicate the Infocom process for making games, even going so far as getting a number of former Infocom employees to help out as testers, but, being a much smaller, decentralized operation, was never able to duplicate the spirited to-and-fro on questions of writing and design that did so much for all of the Infocom Imps’ work. For Steve Meretzky, working hundreds of miles away and accorded a certain untouchable status as Legend’s star designer, the proof of the consequences is in the Spellcasting pudding. I’d like to quote from a recent discussion I had with Bob Bates, in which I asked him his own opinion of the ticking clock and the various other retrograde elements found in the Spellcasting games.

I was never really a fan of a ticking clock. For me personally, the fun of these games comes in the exploration. It comes often not in the solution to the puzzle, but in all the things you discover while you’re trying to solve the puzzle. As a designer, I try to entertain the player during that time, and to encourage the player to try offbeat stuff. Ninety percent of what a player does in an adventure game is wrong. You have to entertain them during that ninety percent in order for them to get the joy of the ten percent that’s actual puzzle-solving. So, I’m not a fan of the time-restricted thing in the same way that I’m not a fan of limiting what the player can carry, making him have to stop and think about swapping things in and out of his inventory. It doesn’t appeal to me as a player or a designer, but other people are different. With Steve, what you’re probably seeing is a designer choice rather than a company choice.

Bob Bates confirmed that Meretzky, as Legend’s star designer, was afforded lots of freedom to make exactly “the game he wanted to make” — freedom which was not afforded to any of the in-house teams who made Legend’s other games, who were expected to hew closely to the more progressive design philosophy Bob outlines above. There’s an important lesson here — important not just in the context of Meretzky’s career, or the history of the adventure game, or even in the context of game design in general. It’s rather an important lesson for all creators, and for those who would nurture them.

I would submit that allowing Meretzky to make exactly the games he wanted to make was in the end greatly to said games’ detriment. I say this not because Steve Meretzky was an egomaniac, was unusually headstrong, or was ever anything other than a very nice, funny, personable, honest fellow who positively oozed with creativity. I say it because Steve Meretzky was human, and creative humans need other humans to push back against them from time to time. I suspect that many of you could name an author or two who used to write taut, compelling books, until they got big enough that no one felt empowered to edit them anymore. Ditto for music; how many once-great bands have tarnished their image with lazy, filler-laden latter-day albums? Game designers are no different from other creative professionals in this respect. Meretzky needed someone to tell him, “Hey, Steve, this aspect of the game is more annoying than fun,” or “Hey, Steve, this huge text dump doesn’t really say much of anything interesting or funny,” or “Hey, Steve, can’t you do something to make this Ernie twerp a bit less odious?”

If the Spellcasting games on the one hand reflect the lack of a sufficiently rigorous design process, on the other they reflect a certain failure of ambition on the part of their creator. Steve Meretzky was a very practical guy at heart, and seems to have taken a strong lesson from the commercial disappointment that had been A Mind Forever Voyaging. He continued to dream grander dreams than the likes of Ernie Eaglebeak and Sorcerer University, but he never pushed hard enough to make his dreams into actualities. The legendary lost Meretzky game, which he broached so many times that it’s become a sort of in-joke among a whole swathe of industry old-timers, was an historical epic taking place on the Titanic‘s one and only voyage. Bob Bates admits that he didn’t want Meretzky to make this game for Legend. Since, as he put it, “everyone knows how that story ends,” he believed there was little commercial appeal to the idea, an opinion which is cast in a very questionable light by the massive success of the James Cameron movie on the same subject of a few years later. What might have been had Mereztky stuck to his guns and brazened out the chance to make this grander vision? It strikes me as unlikely that anyone today would be saying, “Gee, that Titanic game was okay, but I sure would have liked some more Ernie Eaglebeak instead.”

The irony of the Spellcasting series is that these dissipated games did as much to dissipate whatever commercial momentum the newly independent Meretzky had as surely as might have a more ambitious concept. Sales of the Spellcasting games dropped off markedly after the first sold more than 50,000 copies as Legend’s very first product. The dwindling sales doubtless constitute much of the reason that a planned fourth game, Spellcasting 401: The Graduation Ball, was never made. Few regretted its absence overmuch. Between the Spellcasting games and the downright embarrassing Leather Goddesses of Phobos 2, a half-assed graphic adventure he designed for the bankrupt Activision, Meretzky began his career as an independent game designer by ghettoizing himself as the “bad boy of adventure gaming” for years to come.

Ah, well… might-have-beens will always bloom eternal, won’t they? The first Spellcasting game at least can be a lot of fun, and perhaps comes off worse than is really fair here when I place it in the context of the increasingly dispiriting series as a whole. And I do want to note that I uniformly enjoy all of the other Legend text adventures much more than I do Spellcasting 201 and Spellcasting 301. I look forward to having more positive articles to write about them in the future.

The Legend games must still be played as they were originally delivered, as standalone programs running on an old MS-DOS computer or in an MS-DOS emulator, while the games of Infocom and many other earlier text-adventure publishers are much easier to get running thanks to modern interpreters for same. This relative inaccessibility has done much to keep the Legend games from being played as much as they deserve to be. While Bob Bates no longer owns the rights to the Legend games and thus cannot give official permission to host them, he has said that he doesn’t personally object if I do so. So, feel free to download Spellcasting 101Spellcasting 201, and Spellcasting 301 unless and until a scary legal person says it isn’t allowed. Included in each zip files are the associated documentation/feelies and a DOSBox configuration that should work with that particular game, along with a brief note on how to get it running on your system; whether you’re running Windows, OS X, or Linux, it’s all basically the same. I hope this will make it as painless as possible to experience these heretofore hard-to-find and hard-to-get-running pieces of interactive-fiction history.

So, by all means, have a go, starting with Spellcasting 101. Who knows, maybe you’ll find I’m just a boring old stick-in-the-mud and these games are the most hilarious and delightful things ever. Stranger things have happened.


Tags: , ,

Thaumistry: In Charm’s Way

I’d like to bring some exciting news to your attention today. Bob Bates, author of classic games for Infocom and Legend Entertainment and a great friend of this blog, is Kickstarting a new text adventure.

I played the game last year in its alpha state. Bob is very much aware of the ways which text adventures have evolved since the days of Infocom and Legend, and has come up with a great blending of modern and classic here. It’s funny, fun, and occasionally challenging, but always in the right ways. And it’s written with TADS 3, so it’s as technically sophisticated as one could ask for.

I hope some of you will want to join me in helping him to get it polished up and out the door for a variety of platforms. You can learn more on the game’s Kickstarter page or on its home page. Welcome back, Bob!


A Time of Beginnings: Legend Entertainment (or, Bob and Mike’s Excellent Adventure-Game Company)

As the dust settled and the shock faded in the months that followed the shuttering of Infocom, most of the people who had worked there found they were able to convince themselves that they were happy it was finally over, relieved that a clean sharp break had been made. Sure, they had greeted the initial bombshell that the jig was finally up with plenty of disbelief, anger, and sadness, but the fact remained that the eighteen months before that fateful day, during which they had watched their company lose its old swagger, its very sense of itself, had been if anything even more heartbreaking. And yes, there would be plenty of second-guessing among them in the years to come about what might have been if Cornerstone had never existed or if Bruce Davis hadn’t taken over control of Mediagenic,1 but Infocom’s story did nevertheless feel like it had run its natural course, leaving behind something that all of the Bruce Davises in the world could never take away: that stellar 35-game catalog, unmatched by any game developer of Infocom’s era or any era since in literacy, thoughtfulness, and relentless creative experimentation. With that along with all of their fine memories of life inside Infocom’s offices to buoy them, the former employees could move on to the proverbial next chapter in life feeling pretty good about themselves, regarding their time at Infocom as, as historian Graham Nelson so memorably put it, “a summer romance” that had simply been too golden to stay any longer.

Yet there was at least one figure associated with Infocom who was more inclined to rage against the dying of the light than to go gentle into that good night. Bob Bates had come to the job of making text adventures, a job he enjoyed more than anything else he had ever done, just a little bit too late to share the sense of closure felt by the rest of Infocom. Which isn’t to say he hadn’t managed to accomplish anything in the field: Bob had formed a company to challenge Infocom — a company named, appropriately enough, Challenge — that wound up joining them as the only outside developer ever allowed to copy Infocom’s in-house tools and make games for them under contract. Still, it had all happened very late in the day. When all was said and done, he had the dubious distinction of having made the last all-text Infocom game ever, followed by their very last game of all. His summer romance, in other words, had started in the last week of the season, and he’d barely gotten past first base. When he returned from Cambridge, Massachusetts, to his home near Washington, D.C., on the stormy evening of May 5, 1989, having just been informed by Infocom’s head Joe Ybarra that Infocom’s Cambridge offices were being closed and Challenge’s services wouldn’t be needed anymore, his brief life in text adventures just felt so incomplete. And then, he found his roof was leaking. Of course it was.

Some of Bob’s restless dissatisfaction must have come across when, after that unhappy weekend was in the past, he called up Mike Verdu to tell him he would no longer be able to employ Mark Poesch and Duane Beck, the two programmers Mike’s company had sent out to work with Challenge. A remarkably young executive even in a field that has always favored the young, Mike was only in his mid-twenties, but already had an impressive CV.  In his second year at university, he had dropped out to form a consulting company he named Paragon Systems, which had come to employ Poesch and Beck. The two had been sent to Challenge when Bob came calling on Paragon, looking for help programming the games he had just signed a contract with Infocom to create. During the period when Challenge was making games for Infocom, Mike had sold Paragon to American Systems Corporation, a computer-integration firm that did significant business with the Department of Defense. He had stayed on thereafter with the bigger company as director of one of their departments, and Poesch and Beck had continued to work with Challenge, albeit under the auspices of ASC rather than Paragon. But now that would all be coming to an end; thus Bob’s phone call to Mike to inform him that he would have to terminate Challenge’s arrangement with ASC.

In truth, Bob and Mike didn’t know each other all that well prior to this conversation. Mike had always loved games, and had loved having a game company as a client, but Challenge had always had to remain for him as, as he puts it today, “one of many.” The call that Bob now made to Mike therefore began more as a simple transaction between customer and service provider than as a shared commiseration over the downfall of Bob’s business. Still, something that Bob said must have sparked Mike’s interest. The call continued far longer than it ought to have, and soon multiplied into many more conversations. In fairly short order, the conversations led to a suggestion from Mike: let’s start a new company to make and publish text adventures in the Infocom tradition. He even believed he could convince ASC to put up the bulk of the funding for such a company.

Set aside the fact that text adventures were allegedly dying, and the timing was oddly perfect. In 1989 the dominoes were toppling all over Eastern Europe, the four-decade Cold War coming to an end with a suddenness no one could have dreamed of just a few years before. Among the few people in the West not thoroughly delighted with recent turns of events were those at companies like ASC, who were deeply involved with the Department of Defense and thus had reason to fear the “peace dividend” that must lead to budget cuts for their main client and cancelled contracts for them. ASC was eager to diversify to replace the income the budget cuts would cost them; they were making lots of small investments in lots of different industries. In light of the current situation, an investment in a computer-game company didn’t seem as outlandish as it might have a year or two before, when the Reagan defense buildup was still booming, or, for that matter, might have just a year or two later, when the Gulf War would be demonstrating that the American military would not be idle on the post-Cold War world stage.

Though they had a very motivated potential investor, the plan Bob and Mike were contemplating might seem on the face of it counter-intuitive if not hopeless to those of you who are regular readers of this blog. As I’ve spent much time describing in previous articles, the text adventure had been in commercial decline since 1985. That very spring of 1989 when Bob and Mike were starting to talk, what seemed like it had to be the final axe had fallen on the genre when Level 9 had announced they were getting out of the text-adventure business, Magnetic Scrolls had been dropped by their publisher Rainbird, and of course Infocom had been shuttered by their corporate parent Mediagenic. Yet Bob and Mike proposed to fly in the face of that gale-force wind by starting a brand new company to make text adventures. What the hell were they thinking?

I was curious enough about the answer to that question that I made it a point to ask it to both Bob and Mike when I talked to them recently. Their answers were interesting enough, and said enough about the abiding love each had and, indeed, still has for the genre of adventures in text that I want to give each of them a chance to speak for himself here. First, Mike Verdu:

I believed that there was a very hardcore niche market that would always love this type of experience. We made a bet that that niche was large enough to support a small company dedicated to serving it. The genre was amazing; it was the closest thing to the promise of combining literature and technology. The free-form interaction a player could have with the game was a magical thing. There’s just nothing else like it. So, it didn’t seem like a dying art form to me. It just seemed that there were these bigger companies that the market couldn’t support that were collapsing, and that there was room for a smart niche player that had no illusions about the market but could serve that market directly.

I will say that when Bob and I were looking for publishing partners, and went to some trade conferences — through Bob’s connections we were able to meet people like Ken and Roberta Williams and various other luminaries in the field at the time — everybody said, “You have no idea what you’re doing. The worst idea in the world is to start a game company. It’s the best way to take a big pile of money and turn it into a small pile of money. Stay away!” But Bob and I are both stubborn, and we didn’t listen.

Understanding your market opportunity is really key when you’re forming a company. With Legend, we were very clear-eyed about the fact that we were starting a small company to serve a small market. We didn’t think it would grow to be a thousand people or take over the world or sell a million units of entertainment software per year. We thought there was this amazing, passionate audience that we could serve with these lovingly crafted products, and that would be very fulfilling creatively. If you’re a creative person, I think you have to define how big the audience is that is going to make you feel fulfilled. Bob and I didn’t necessarily have aspirations to reach millions of people. We wanted to reach enough people that we could make our company viable, make a living, and create these products that we loved.

And Bob Bates:

We recognized the risk, but basically we just still believed in the uniqueness of the parser-driven experience — in the pleasure and the joy of the parser-driven experience. By then, there were no other major parser-driven games around, and we felt that point-and-click was a qualitatively different experience. It was fun, but it was different. It was restrictive in terms of what the player could do, and there was a sense of the game world closing in on you, that you could only do what could be shown. Brian Moriarty had a great quote that I don’t remember exactly, but it was something like “you can only implement what you can afford to show, and you can’t afford to show anything.” As a player, I loved the freedom to input whatever I wanted, and I loved the low cost of producing that [form of interaction]. If there’s an interesting input or interaction, and I can address it in a paragraph of text, that’s so much cheaper than having an artist spend a week drawing it. Text is cheap, so we felt we could create games economically. We felt that competition in that niche wasn’t there anymore, and that it was a fun experience that there was still a market for.

Reading between the lines just a bit here, we have a point of view that would paint the failure of Infocom more as the result of a growing mismatch between a company and its market than as an indication that it was genuinely impossible to still make a living selling text adventures. Until 1985, the fulcrum year of the company’s history, Infocom had been as mainstream as computer-game publishers got, often placing three, four, or even five titles in the overall industry top-ten sales lists each month. Their numbers had fallen off badly after that, but by 1987 they had stabilized to create a “20,000 Club”: most games released that year sold a little more or less than 20,000 copies. Taking into account the reality that every title would never appeal enough to every fan to prompt a purchase — especially given the pace at which Infocom was pumping out games that year — that meant there were perhaps 30,000 to 40,000 loyal Infocom fans who had never given up on the company or the genre. Even the shrunken Infocom of the company’s final eighteen months was too big to make a profit serving that market, which was in any case nothing Bruce Davis of Mediagenic, fixated on the mainstream as he was, had any real interest in trying to serve. A much smaller company, however, with far fewer people on the payroll and a willingness to lower its commercial expectations, might just survive and even modestly thrive there. And who knew, if they made their games really well, they might just collect another 30,000 or 40,000 new fans to join the Infocom old guard.

This wasn’t to say that Bob and Mike could afford to return to the pure text that had sufficed for 31 of Infocom’s 35 adventure games. To have any chance of attracting new players, and quite possibly to have any chance of retaining even the old Infocom fans, they were well aware that some concessions to the realities of the contemporary marketplace would have to be made. Their games would include an illustration for every location along with occasional additional graphics, sound effects, and music to break up their walls of text. Their games would, in other words, enhance the Infocom experience to suit the changing times rather than merely clone it.

In the same spirit of maximizing their text adventures’ contemporary commercial potential, they very early on secured the services of Steve Meretzky, Infocom’s single most well-known former Implementor, who had worked on some of the company’s most iconic and successful titles. With Meretzky’s first game for their company, Bob and Mike would try to capitalize on his reputation as the “bad boy of adventure gaming” — a reputation he enjoyed despite the fact that he had only written one naughty adventure game in his career to date. Nevertheless, Bob encouraged Meretzky to “take the gloves off,” to go much further than he had even in his previous naughty game Leather Goddesses of Phobos. Meretzky’s vision for his new game can perhaps be best described today as “Animal House meets Harry Potter” (although, it should be noted, this was many years before the latter was published). It would be the story of a loser who goes off to Sorcerer University to learn the art and science of magic, whilst trying his best to score with chicks along the way. Of course, this being an adventure game, he would eventually have to save the world as well, but the real point was the spells and the chicks. The former would let Meretzky revisit one of the most entertaining puzzle paradigms Infocom had ever devised: the Enchanter series’s spell book full of bizarre incantations that prove useful in all sorts of unexpected ways. The latter would give Bob and Mike a chance to prove one more time the timeless thesis that Sex Sells.

So, the Meretzky game seemed about as good as things could get as a commercially safe bet, given the state of text adventures in general circa 1989. Meanwhile, for those players less eager to be titillated, Bob Bates himself would make what he describes today as a “classical” adventure, a more sober-minded time-travel epic full of intricately interconnected puzzles and environments. Between the two, they would hopefully have covered most of what people had liked about the various games of Infocom. And the really hardcore Infocom fans, of course, would hopefully buy them both.

In making their pitch to ASC and other potential investors, Bob and Mike felt ethically obligated to make careful note of the seeming headwinds into which their new company would be sailing. But in the end ASC was hugely eager to diversify, and the investment that was being asked of them was relatively small in the context of ASC’s budget. Bob and Mike founded their company on about $500,000, the majority of which was provided by ASC, alongside a handful of smaller investments from friends and family. (Those with a stake in Bob’s old company Challenge also saw it rolled over into the new company.) ASC would play a huge role during this formative period, up to and including providing the office space out of which the first games would be developed.

An ASC press release dated January 8, 1990, captures the venture, called GameWorks at the time, at this embryonic stage of high hopes and high uncertainty. Bob Bates is quoted as saying that “GameWorks products combine the best of several existing technologies in an exciting new format,” while Mike Verdu, who would remain in his old role at ASC in addition to his new one as a software entrepreneur for another couple of years, says that “ASC’s interest in this venture stems from more than just making money over the short term. The goal is to establish a self-sustaining software-publishing company.” Shortly after this press release, the name of said company would be changed from GameWorks to Legend Entertainment, harking back to the pitch for an “Immortal Legends” series of games that had first won Bob a contract with Infocom.

The part of the press release that described GameWorks/Legend as a “software-publishing company” was an important stipulation. Mike Verdu:

I remember making these spreadsheets early on, trying to understand how companies made money in this business. It became very clear to me very quickly that life as an independent developer, without the publishing, was very tough. You scrambled for advances, and the royalties you got off a game would never pay for the advances unless you had a huge hit. Your destiny was so tied to the publisher, to the vagaries of the producer that might get assigned to your title, that it just was not an appealing path at all.

In a very fundamental way, Legend needed to be a publisher as well as a developer if they were to bring their vision of text adventures in the 1990s to fruition. It was highly doubtful whether any of the other publishers would be willing to bother with the niche market for text adventures at all when there were so many other genres with seemingly so much greater commercial potential. In addition, Bob and Mike knew that they needed to have complete control of their products, from the exact games they chose to make to the way those games were packaged and presented on store shelves. They recognized that another part of becoming the implicit successor to Infocom must be trying as much as possible to match the famous Infocom packaging, with the included “feelies” that added so much texture and verisimilitude to their interactive fictions. One of the most heartbreaking signs of Infocom’s slow decline, for fans and employees alike, had been the gradual degradation of their games’ physical presentation, as the cost-cutters in Mediagenic’s Silicon Valley offices took away more and more control from the folks in Cambridge. Bob and Mike couldn’t afford to have their company under a publisher’s thumb in similar fashion. At the same time, though, a tiny company like theirs was in no position to set up its own nationwide distribution from warehouse to retail.

It was for small publishers facing exactly this conundrum that Electronic Arts and Mediagenic during the mid-1980s had pioneered the concept of the “affiliated label.” An affiliated label was a small publisher that printed their own name on their boxes, but piggy-backed — for a fee, of course — on the network of a larger publisher for distribution. By the turn of the decade, the American computer-games industry as a whole had organized itself into eight or so major publishers, each with an affiliated-label program of one stripe or another of its own, with at least several dozen more minor publishers taking advantage of the programs. As we’ve seen in other articles, affiliated-label deals were massive potential minefields that many a naive small publisher blundered into and never escaped. Nevertheless, Legend had little choice but to seek one for themselves. Thanks to Mike Verdu’s research, they would at least go in with eyes open to the risk, although nothing they could do could truly immunize them from it.

In seeking a distribution deal, Legend wasn’t just evaluating potential partners; said partners were also evaluating them, trying to judge whether they could sell enough games to make a profitable arrangement for both parties. This process, like so much else, was inevitably complicated by Legend’s determination to defy all of the conventional wisdom and continue making text adventures — yes, text adventures with graphics and sound, but still text adventures at bottom. And yet as Bob and Mike made the rounds of the industry’s biggest players they generally weren’t greeted with the incredulity, much less mockery, one might initially imagine. Even many of the most pragmatic of gaming executives felt keenly at some visceral level the loss of Infocom, whose respect among their peers had never really faded in tune with their sales figures — who, one might even say, had had a certain ennobling effect on their industry as a whole. So, the big players were often surprisingly sympathetic to Legend’s cause. Whether such sentiments could lead to a signature on the bottom line of a contract was, however, a different matter entirely. Most of the people who had managed to survive in this notoriously volatile industry to this point had long since learned that idealism only gets you so far.

For some time, it looked like a deal would come together with Sierra. Ken Williams, who never lacked for ambition, was trying to position his company to own the field of interactive storytelling as a whole. Text adventures looked destined to be a very small piece of that pie at best in the future, but that piece was nevertheless quite possibly one worth scarfing up. If Sierra distributed Legend’s games and they proved unexpectedly successful, an acquisition might even be in the cards. Yet somehow a deal just never seemed to get done. Mike Verdu:

There seemed to be genuine interest [at Sierra], but it was sort of like Zeno’s Paradox: we’d get halfway to something, and then close that distance by half, and then close that distance by half, and nothing ever actually happened. It was enormously frustrating — and I never could put my finger on quite why, because there seemed to be this alignment of interests, and we all liked each other. There was always a sense of a lot of momentum at the start. Then the momentum gradually died away, and you could never actually get anything done. Now that I’ve become a little more sophisticated about business, that suggests to me that Ken was probably running around trying to make a whole bunch of things happen, and somebody inside his company was being the sort of check and balance to his wanting to do lots and lots of stuff. There were probably a lot of things that died on the vine inside that company.

Instead of Sierra, Legend wound up signing a distribution contract with MicroProse, who were moving further and further from their roots in military simulations and wargames in a bid to become a major presence in many genres of entertainment software. Still, “Wild Bill” Stealey, MicroProse’s flamboyant chief, had little personal interest in the types of games Legend proposed to make or the niche market they proposed to serve. Mike Verdu characterizes Sierra’s interest as “strategic,” while MicroProse’s was merely “convenient,” a way to potentially boost their revenue picture a bit and offset a venture into standup-arcade games that was starting to look like a financial disaster. MicroProse hardly made for the partner of Legend’s dreams, but needs must. Wild Bill was willing to sign where Ken Williams apparently wasn’t.

In the midst of all these efforts to set up the infrastructure for a software-publishing business, there was also the need to create the actual software they would publish. Bob Bates’s time-travel game fell onto the back-burner, a victim of the limited resources to hand and the fact that so much of its designer’s time was being monopolized by practical questions of business. But not so Steve Meretzky’s game. As was his wont, Meretzky had worked quickly and efficiently from his home in Massachusetts to crank out his design. Legend’s two-man programming team, consisting still of the Challenge veterans Duane Beck and Mark Poesch, was soon hard at work alongside contracted outside artists and composers to bring Spellcasting 101: Sorcerers Get All the Girls, now planned as Legend’s sole release of 1990, to life in all its audiovisual splendor.

Setting aside for the moment all those planned audiovisual enhancements, just creating a reasonable facsimile of the core Infocom experience presented a daunting challenge. Throughout Infocom’s lifespan, from the 1980 release of Zork I through Bob Bates’s own 1989 Infocom game Arthur: The Quest for Excalibur, no other company had ever quite managed to do what Legend was now attempting to do: to create a parser as good as that of Infocom. Legend did have an advantage over most of Infocom’s earlier would-be challengers in that they were planning to target their games to relatively powerful machines with fast processors and at least 512 K of memory. The days of trying to squeeze games into 64 K or less were over, as were the complications of coding to a cross-platform virtual machine; seeing where the American market was going, Legend planned to initially release their games only for MS-DOS systems, with ports to other platforms left only as a vague possibility if one of their titles should prove really successful. Both the Legend engine and the games that would be made using it were written in MS-DOS-native C code instead of a customized adventure programming language like Infocom’s ZIL, a decision that also changed the nature of authoring a Legend game in comparison to an Infocom game. Legend’s designers would program many of the simpler parts of their games’ logic themselves using their fairly rudimentary knowledge of C, but would always rely on the “real” programmers for the heavy lifting.

But of course none of these technical differences were the sort of things that end users would notice. For precisely this reason, Bob Bates was deeply worried about the legal pitfalls that might lie in attempting to duplicate the Infocom experience so closely from their perspective. The hard fact was that he, along with his two programmers, knew an awful lot about Infocom’s technology, having authored two complete games using it, while Steve Meretzky, who had authored or coauthored no less than seven games for Infocom, knew it if anything even better. Bob worried that Mediagenic might elect to sue Legend for theft of trade secrets — a worry that, given the general litigiousness of Mediagenic’s head Bruce Davis, strikes me as eminently justified. To address the danger, Legend elected to employ the legal stratagem of the black box. Bob sat down and wrote out a complete specification for Legend’s parser-to-be. (“It was a pretty arcane, pretty strange exercise to do that,” he remembers.) Legend then gave this specification for implementation to a third-party company called Key Systems who had never seen any of Infocom’s technology. “What came back,” Bob says, “became the heart of the Legend engine. Mark and Duane then built additional functionality upon that.” The unsung creators of the Legend parser did their job remarkably well. It became the first ever not to notably fall down anywhere in comparison to the Infocom parser. Mediagenic, who had serious problems of their own monopolizing their attention around this time, never did come calling, but better safe than sorry.

The Legend Interface in a Nutshell

A game can be played in one of three modes. This one, the default, is the most elaborate — not to say cluttered. Note the long menus of verbs — 120 (!) of them, with a commonly used subset thankfully listed first — and nouns to the left. (And don’t worry, this area from Spellcasting 101 is a “fake” maze, not a real one.)

A second mode, which I suspect was the most commonly used by real players in the wild, removes the command-entry menus in favor of allowing more space for the text window, but retains the compass rose and illustrations.

Finally, strict adherents to the ethos of text-and-only-text can indeed play the game as a text-only adventure. The existence of significant numbers of such purists was probably more theoretical than actual, but Legend accommodated them nevertheless.

By tapping the function keys, you can replace the illustration with the current room description or your current inventory without having to burn a turn on the task.

Or you can show a map where the picture usually lives.


Anxious to make their games as accessible as possible despite their equally abiding determination to become the implicit heir to Infocom, Legend designed for their new engine a menu-based system for inputting commands that could serve as an alternative to typing them in. Bob Bates, the mastermind behind the system:

One of the biggest barriers to text adventures at the time was that people didn’t know how to type. I knew how to type only because the principal of my high school forced me in my sophomore year to take a typing class instead of a third language. At the time, typing was for girls; men didn’t type. It was a barrier for players.

So, we said that we need an interface that will let somebody play using only the mouse. This was a huge problem. How do you do that without giving too much away? One day as I was pondering this, I realized that once you select a verb you don’t need another verb. So, the menu that contains verbs can go away. You’re looking at a list of verb/noun [combinations]: “get box,” “kick wall.” But if you want a sentence with a preposition, once you’ve clicked on the verb you don’t need another verb, so you can replace that first [verb] list with prepositions — and not only that, but prepositions that are only appropriate to that verb. That was an actual insight; that was a cool idea.

The menu on the left had the twenty or so most common verbs first, but underneath that, going down in alphabetical order, was a list of many, many more verbs. You could scroll down in that list, and it might actually suggest interactions you hadn’t thought of. Basically it preserved the openness of the interaction, but avoided the other big bugaboo of parser-driven games: when the parser will come back and say, “I don’t understand that.” With this system, that could never happen. And that was, I thought, huge. Everything was there in front of you if you could figure out what to do. [Parsing errors] became a thing of the past if you wanted to play in that mode.

Then of course we had full-screen text mode if you wanted to play that way, and we had a sort of hybrid half-and-half mode where there was parser-driven text across the bottom, but you still had graphics at the top. I thought it was important that players could play the game the way they wanted to, and I thought it added to the experience by taking away two of the big problems. One was people who didn’t like to type or couldn’t type or were two-finger typists. Number two was when you would type a whole command and there was an error in the first word; the parser says, “I don’t know that word,” and you have to type the whole command again. That interface took away that pain in the ass.

While Bob’s points are well taken, particularly with regard to the lack of typing skills among so much of the general public at the time, the Legend menu-based interface looks very much of its time today. Having the menu appear onscreen by default has had the unfortunate side-effect of making the Legend games look rather cluttered and ugly in contrast to the Infocom classics, with their timeless text-only approach. That does a real disservice to the games hidden inside the Legend interface, which often stand up very well next to many of the works of Infocom.

Aesthetics aside, I remain skeptical of the real long-term utility of these sorts of interfaces in general, all the rage though they were during the twilight of the text adventure’s commercial era. Certainly there must come a point where picking through a list of dozens of verbs becomes as confusing as trying to divine the correct one from whole cloth. A better solution to the guess-the-verb problem is to create a better parser — and, to be fair, Legend games give no ground for complaint on that score; text-adventure veteran though I admittedly am, I can’t recall ever struggling to express what I was trying to do to a Legend game. The problem of correcting typos without having to type the entire command again, meanwhile, could have been efficiently addressed by including a command-history buffer that the player could navigate using the arrow keys. The omission of such a feature strikes me as rather inexplicable given that the British company Level 9 had begun to include it in their games as far back as 1986.

Although I don’t believe any serious surveys were ever made, it would surprise me if most Legend players stuck with the menu-based interface for very long once they settled down to play. “I played the game this way for fifteen minutes before deciding to bag it and type in all my commands,” wrote one contemporary Spellcasting 101 reviewer who strikes me as likely typical. “For me, this was quicker.” “Frankly, I find the menu to be of little use except to suggest possible commands in tough puzzle situations,” wrote another. Even Steve Meretzky, the author of Legend’s first game, wasn’t a fan:

The impetus for the interface was not a particular feeling that this was a good/useful/friendly/clever interface, but rather a feeling that text adventures were dying, that people wanted pictures on the screen at all times, and that people hated to type. I never liked the interface that much. The graphic part of the picture was pretty nice, allowing you to move around just by double-clicking on doors in the picture, or pick things up by double-clicking on them. But I didn’t care for the menus for a number of reasons. One, they were way more kludgey and time-consuming than just typing inputs. Two, they were giveaways because they gave you a list of all possible verbs and all visible objects. Three, they were a lot of extra work in implementing the game, for little extra benefit. And four, they precluded any puzzles which involved referring to non-visible objects.

Like Meretzky, I find other aspects of the Legend engine much more useful than the menu-based command interface. In the overall baroque-text-adventure-interface sweepstakes, Magnetic Scrolls’s Magnetic Windows-based system has the edge in features and refinement, but the Legend engine does show a real awareness of how real players played these types of games, and gives some very welcome options for making that experience a little less frustrating. The automap, while perhaps not always quite enough to replace pen and paper (or, today, Trizbort), is nevertheless handy, and the ability to pull up the current room description or your current inventory without wasting a turn and scrolling a bunch of other text away is a godsend, especially given that there’s no scrollback integrated into the text window.

The graphics and music in the Legend games still hold up fairly well as well, adding that little bit of extra sizzle. (The occasional digitized sound effects, on the other hand, have aged rather less well.) Right from the beginning with Spellcasting 101, Legend proved willing to push well beyond the model of earlier, more static illustrated text adventures, adding animated opening and closing sequences, interstitial graphics in the chapter breaks, etc. It’s almost enough to make you forget at times that you’re playing a text adventure at all — which was, one has to suspect, at least partially the intention. Certainly it pushes well beyond what Infocom managed to do in their last few games. Indeed, I’m not sure that anyone since Legend has ever tried quite so earnestly to make a real multimedia production out of a parser-based game. It can make for an odd fit at times, but it can be a lot of fun as well.

Spellcasting 101 was released in October of 1990, thereby bringing to a fruition the almost eighteen months of effort that had followed that fateful Cinco de Mayo when Bob Bates had learned that Infocom was going away. I plan to discuss the merits and demerits owed to Spellcasting 101 as a piece of game design in my next article. For now, it should suffice to say that the game and the company that had produced it were greeted with gushing enthusiasm by the very niche they had hoped to reach. Both were hailed as the natural heirs to the Infocom legacy, carrying the torch for a type of game most had thought had disappeared from store shelves forever. Questbusters magazine called Spellcasting 101 the “Son of Infocom” in their review’s headline; the reviewer went on to write that “what struck me most about the game is that it is exactly as I would have expected Infocom games to be if the company was still together and the veteran designers were still working in the industry. I kid you not when I say to watch Legend over the years.” “It’s such a treat to play an Infocom adventure again,” wrote the adventuring fanzine SynTax. “I know it isn’t an Infocom game as such, but I can’t help thinking of it as that.”

This late in the day for the commercial text adventure, it was these small adventure-centric publications, along with the adventure-game columnists for the bigger magazines, who were bound to be the most enthusiastic. Nevertheless, Spellcasting 101 succeeded in proving the thesis on which Bob Bates and Mike Verdu had founded Legend Entertainment: that there were still enough of those enthusiasts out there to support a niche company. In its first six months on the market, Spellcasting 101 sold almost 35,000 units, more than doubling Bob and Mike’s cautious prediction of 16,000 units. By the same point, the Legend hint line had fielded over 35,000 calls. For now — and it would admittedly be just for a little while longer — people were buying and, as the hint-line calls so amply demonstrated, playing a text adventure again in reasonable numbers, all thanks to the efforts of two men who loved the genre and couldn’t quite let it go.

A “Presentation to Stockholders and Directors” of Legend from May of 1991 provides, like the earlier ASC press release, another fascinating real-time glimpse of a business being born. At this point Timequest, Bob Bates’s “classical” time-travel adventure, is about to be released at last, Spellcasting 201 is already nearing completion, and a first licensed game is in the offing, to be based on Frederick Pohl’s Gateway series of science-fiction novels. “MicroProse has done an outstanding job of selling and distributing the product,” notes the report, but “has been less than responsive on the financial side of the house. Our financial condition is precarious. We spent most of the Spellcasting 101 revenues in development of Timequest. We are living hand to mouth. We have come a long way and we are building a viable business, but the costs were greater than expected and the going has often been rough.”

Rough going and living hand to mouth were things that Legend would largely just have to get used to. The games industry could be a brutal place, and a tiny niche publisher like Legend was all but foreordained to exist under a perpetual cloud of existential risk. Still, in return for facing the risk they were getting to make the games they loved, and giving the commercial text adventure a coda absolutely no one had seen coming on that unhappy day back in May of 1989. “We did more things right than we did wrong,” concludes the May 1991 report. “This is a workable definition of survival.” Survival may have been about the best they could hope for — but, then again, survival is survival.

(Sources: Questbusters of March 1991; SynTax Issue 11; Computer Gaming World of November 1990 and March 1991; the book Game Design Theory and Practice by Richard Rouse III; Bob Bates’s interview for Jason Scott’s Get Lamp documentary, which Jason was kind of enough to share with me in its entirety. But the vast majority of this article is drawn my interviews with Bob Bates and Mike Verdu; the former dug up the documents mentioned in the article as well. My heartfelt thanks to both for making the time to talk with me and to answer my many nitpicky questions about events of more than 25 years ago.)

  1. Mediagenic was known as Activision until mid-1988. To avoid confusion, I just stick with the name “Mediagenic” in this article. 


Tags: , , ,