The Short, Happy Life of Infocom

Infocom, king of the text adventure and the first behemoth of American computer game development, began not with a bang, but with an internet meme.

Long before “All Your Base” and the “O R’LY? owl,” there was Adventure. Released in 1975 onto the ARPANet (the internet’s predecessor), Will Crowther’s Adventure was essentially a simulation of a caving expedition he’d done in Kentucky. But his co-workers loved it, and they passed it along to their friends. As the game traveled across networks, it was revised and re-written, particularly when Stanford researcher Don Woods beefed up the storyline and added some Tolkien-esque flair. By the time it migrated onto MIT’s mainframe in 1976, the game had blossomed into somewhat of a pre-internet net phenomenon.

As a game, however, Adventure was far from perfect. Aside from programming artifacts and design flaws, Adventure‘s biggest drawback was its parser (the program that translated a player’s input into directions the game could understand). The parser could only handle two-word commands, like “go north” or “take sword,” frustrating players who wanted more natural, complex commands.

So, in 1977, four MIT students – Marc Blank, Tim Anderson, Dave Lebling and Bruce Daniels – decided to make their own game in Adventure’s image. Their version retained Adventure‘s basic interface but featured an improved parser, which now could understand complete sentences. The new game, which they named Zork, appeared on the school’s mainframe in 1979, where anyone who had access could play. Like Adventure before it, Zork steadily earned a worldwide cult following.

***

Shortly after Zork‘s publication, a few MIT computer science students – including Zork‘s writers – decided that they wanted to work together outside a stuffy university setting. That summer, they started their own software company and called it Infocom.

In retrospect, the idea was painfully naïve. Although Infocom’s founders vaguely knew they wanted to create business software, they had no model, no business plan, not even a product. But as luck would have it, the home PC market had just entered its first boom. Blank, along with his friend, Joel Berez, suggested bringing Zork to this new home audience; the game’s profits could help finance Infocom until its business division took off.

But Zork was monstrous; the game’s memory requirements alone made most contemporary PCs weep in horror. To make Zork run on a weaker system would require substantial re-engineering.

Blank and Berez started their ambitious task by first constructing the “Z-machine,” a virtual machine that included only the operations Zork specifically required. The Z-machine would also leave most of the game code on disk, loading sections into the system’s memory only when necessary. Next, they built a new compiler, the Z-machine Interpretive Program (ZIP), which ported the game to a given operating system. To run Zork, all a computer needed was its platform-specific ZIP. With the variety of computer models available at the time, this portability would become the crux of Infocom’s success.

But even after re-tooling, Zork was still too big. So, the game was divided into three parts, and the first segment was released in 1980 as Zork I.

Although Zork I would eventually sell more than a million copies over several platforms, initial sales were slow. This was partly because Zork‘s publisher, Personal Software, Inc., incorrectly marketed the game as a hack-and-slash: The cover art featured a mustachioed barbarian with a gleaming sword, vanquishing a cowering orc. But when Personal Software dropped Infocom in 1981, Infocom brought its publishing in-house and started over. They discarded all of Zork‘s original packaging and made their own much-improved game materials.

In-house publishing was a brilliant move for Infocom, since it allowed them more creative freedom. This proved especially true in 1982, when Marc Blank, working on his detective mystery, Deadline, realized he couldn’t fit everything he wanted into the actual game. So, he designed extra items to include in the packaging, like photos, lab results and pills. Reviewers and players alike loved it, and thus began Infocom’s famous tradition of “feelies.”

As Zork I steadily gained in popularity, Infocom released the rest of the Zork trilogy, as well as some standalone titles: Deadline, Starcross, Planetfall and others. These games were consistent critical and commercial hits, and soon, Infocom had earned a reputation for enjoyable, well-written games. This goodwill continued for years, and some of their later titles – like Trinity, A Mind Forever Voyaging and The Hitchhiker’s Guide to the Galaxy – would be remembered among the greatest games ever made.

Much of Infocom’s success rested with its employees: Most were young, well-educated and without families. Lacking other commitments to hold them back, the programmers (known as “Implementers,” or Imps for short) could turn games around quickly – within nine months, on average – and for less than $500,000.

Moreover, most Imps enjoyed developing games for Infocom because it was a genuinely fun place to work. There were parties, costume contests and weekly Imps’ Lunches, where employees got together and discussed ideas freely. In the Imps’ Lounge, aborted board games and half-finished diversions covered the tables. Sometimes, Imps would stage hermit crab races on a makeshift tabletop racetrack. The quirky culture and laid-back creativity made Infocom unlike anything the gaming industry had ever seen.

***

Emboldened by their gaming success, Infocom executives felt confident enough to return to their original plan: Developing business software. In 1982, the company created a Business Products division and started work on a relational database called Cornerstone.

At the time, Cornerstone seemed like a bright idea. Business software promised higher profit margins: Whereas the typical game retailed for $50, database software sold for 10 times that. Also, many employees believed the company should diversify in order to survive; fickle gamers could quit their Infocom addiction at any time, but business clients tended to invest heavily and stick with their software.

Yet, from the start, Cornerstone suffered money troubles. Finding investors proved difficult, since people were understandably skeptical of a gaming company interested in making “serious” software. Moreover, Business Products hurled cash at new offices and resources, money that Infocom hadn’t yet made.

This might have resolved itself, had Cornerstone sold well. But when the software was released in 1985, it was a commercial flop. In its first year, Cornerstone only sold 10,000 copies, less than 40% of its projected sales.

Cornerstone’s problem was not poor design but obsolescence. It operated on the same virtual machine technology that had made Zork and other games so portable, but by 1985, the IBM-PC had become the dominant computer platform. Portability was no longer an issue. In fact, Cornerstone’s virtual technology slowed the IBM to a crawl.

Unfortunately, Cornerstone’s failure coincided with a lull in the gaming industry. For the first time ever, Infocom’s game profits stagnated, and Zork I slipped from the number one slot – where it had been for three years – to number 10.

***

Infocom was in dire financial straits. Even three rounds of layoffs and across-the board salary cuts couldn’t save the company. By the end of 1985, it was clear Infocom would need outside assistance.

In 1986, Jim Levy, CEO of Activision (and huge Infocom fan), offered to buy the struggling company. Infocom accepted, and that February, Activision purchased the company for $7.5 million.

In general, most Infocom employees looked favorably upon the union, realizing it was the only way to keep their beloved company afloat. It didn’t hurt that Activision’s corporate culture was similar to Infocom’s, or that Levy promised to stay out of their affairs.

But then, six months later, Activision also ran into financial trouble. Levy was replaced by Bruce Davis, the only board member who’d been against the Infocom merger.

Davis was far more hands-on than his predecessor, and many of his ill-conceived decisions crippled Infocom. Not only did he require the company to use Activision’s packaging plant instead of their own, doubling their marketing costs, he also restructured Infocom’s selling practices. Previously, older Infocom games sold like books, side-by-side with newer ones. Davis halted that, retiring the older ones for good. To fill the now-empty shelves, he ordered Infocom to produce eight games a year, instead of the four or five as before.

To make matters worse, text adventures were no longer selling as well as they used to, especially since graphical adventures had finally become competitive on the market. But Infocom had little time or money to experiment with graphics, and the few such titles they released sold poorly.

By 1989, Activision had had enough. That May, the company laid off most of the remaining Infocom staff and integrated Infocom’s sales, marketing and customer support teams into its own. Infocom was officially dead. While Activision continued to release greatest hit compilation of Infocom games (like The Lost Treasures of Infocom), text adventures themselves faded away.

That is, until they found a home on the internet. Returning to the primordial soup from which they spawned, text adventures have inspired a quiet but substantial net following. Today, you can find most of Infocom’s games online, and Zork is only a short Google away.

Lara Crigger is a freelance science and tech writer whose work on videogames has appeared in Computer Games Magazine and Gamers with Jobs. Her favorite Infocom game is Trinity, but she still has a soft spot for Leather Goddesses Of Phobos.