The History of Rogue: Have @ You, You Deadly Zs

[In the latest in a series of Gamasutra-exclusive bonus material originally to be included in Bill Loguidice and Matt Barton's new book Vintage Games: An Insider Look at the History of Grand Theft Auto, Super Mario, and the Most Influential Games of All Time, the authors investigate the origins, permutations, and legacy of Rogue, one of the most quietly influential and long-lived games ever created. Previously in this 'bonus material' series: Elite, Tony Hawk's Pro Skater, Pinball Construction Set and Pong.]

Rogue: Exploring the Dungeons of Doom (aka Rogue), created in the early 1980s[1] by Michael Toy and Glenn Wichman, is an intriguing game for many reasons. For one, it's still being actively played, ported, enhanced, and forked[2] two decades later -- a fact that challenges its description as just a "vintage" or "retro" game.

It's also among a scant handful of games that have achieved worldwide recognition despite originating on UNIX[3], a platform better suited for science and industry than computer games.

Indeed, many UNIX terminals lack all but the most basic audiovisuals, so Rogue and the majority of other games for the platform have to rely on text or "character-set graphics," which is to say pictures made up of characters like = and [][4].

It's also important to consider the community in which Rogue originated. Programmers like Toy and Wichman had a different background than most developers of commercial games, who were (and are) primarily concerned with making profit.

Whereas developers like Richard Garriott (see book Chapter 23, "Ultima (1980): The Immaculate Conception of the Computer Role-Playing Game") and Roberta Williams (see book Chapter 11, "King's Quest: Quest for the Crown (1984): Perilous Puzzles, Thorny Thrones") would have thought it crazy to share the source code of their bestselling games with their competitors, many UNIX developers (especially those in academic settings) tended to freely share their programs and their code.

Rogue and many of its derivatives rely entirely on character-set graphics to depict dungeons and monsters. Shown here is a battle between the player's character (the "@" symbol) and a kobold (represented by the "K").

The money was assumed to be in the hardware, not the software. Hardware companies modified UNIX as they saw fit, but it soon became apparent that an industry-wide standard was needed to ensure compatibility.

AT&T and Bell Labs, who owned UNIX, began to license closed-source versions of UNIX for commercial use, but other companies banded together to create open-source alternatives. One of these was BSD UNIX, developed at the University of California at Berkeley, where Michael Toy was a student.

One important feature of BSD UNIX was Ken Arnold's curses, a library of functions for drawing pictures on the screen using a terminal's character set. Although there were certainly serious uses for curses, it was also a great tool for making games.

Toy and Wichman, who were schoolmates at the time, had already dabbled in game development before, creating text adventures[5], among other small projects. Curses inspired them to create a "graphical" role-playing game, which would depict dungeons from a top-down perspective.

The monsters, treasures, traps, and other objects would be represented by certain symbols; Z for zombie, for instance. The player's character was an @ -- perhaps a play on "where you're at." However, the graphics routines, while certainly innovative, weren't the game's most distinguishing feature; procedurally generated -- or created on the fly -- dungeons were.

Shown here is Rogue Clone IV, which shows what the game could look like using the IBM PC's character set and basic color.

[1] The BSD UNIX distribution 4.2 in 1980 included a binary version of Rogue, allowing for wide distribution, which is the pivotal release event we'll use for initial dating. See http://users.tkk.fi/~eye/roguelike/rogue.html.

[2] A "fork" occurs when a programming project splits off into separate projects, usually controlled by different groups of programmers.

[3] Also written "Unix".

[4] A similar scenario played out for many home users who utilized the popular CP/M operating system standard of the late 1970s and early 1980s, which had few, if any, graphical features in its myriad implementations. Naturally, CP/M was a target for many roguelike developments, including the commercial Nemesis from SuperSoft (1981).

[5] See book Chapter 25, "Zork: Text Imps versus Graphics Grues" for more on text adventures.

Although the text adventures Toy and Wichman played and developed were fun, they suffered from limited replayability. Once y

ou've solved all the puzzles in Zork, for instance, there's little reason to continue playing. What Toy and Wichman desired to make was a game that would be different each time, never offering the exact same gameplay twice.

Though the game would offer a basic story and goal (fetch the Amulet of Yendor from the bottom of the dungeon), the real fun was exploring the dungeons, vanquishing increasingly ferocious monsters, collecting valuable treasures and equipment, and strengthening one's character.

The control scheme was as intuitive as one might expect from the era before mice and pull-down menus. Besides the basic movement keys (h, j, k, and l), players also had to remember somewhat arbitrary commands like "q" to quaff a potion, or "e" to eat food.

Although the control scheme was relatively easy to master, the game itself was often quite challenging. Sudden death could occur at any moment, particularly if the character weren't well equipped and stocked with potions and scrolls.

Still, though death was common, starting over wasn't so tedious, as the dungeons would be randomized each time. "Every time you played," said Wichman, "you got a new adventure. That's really what made it so popular for all those years in the early eighties."[6]

Toy and Wichman's game was quite popular, but it didn't get its big break until it was added to Version 4.2 of BSD UNIX, the operating system of choice on university mainframes all over the world.

According to Wichman, "over the next three years, Rogue became the undisputed most popular game on college campuses."[7] The game's rousing success among the college crowd seemed to bode well for its commercial potential in the computer games market; after all, Zork's developers had followed a similar path and earned millions.

Front (left) and back (right) of the box for the Epyx version of Rogue, Atari ST version. Despite Epyx's strong distribution channel and advertising resources, Rogue was not considered a commercial success.

Wichman himself wasn't involved in the first effort to market the game commercially. Toy had teamed up with another programmer named Jon Lane, who was able to port the game to the IBM PC.

The two started their own company named A.I. Design and tried to sell the product themselves, but in 1983 called upon Epyx to help market and distribute it as Rogue.[8] It was soon ported to the Apple Macintosh, Atari ST, Commodore Amiga, and Radio Shack Color Computer 3, among others, with each version receiving its own set of enhancements and quirks.

A screenshot from the Atari ST version of Epyx's Rogue. Though far more graphically interesting than most other versions of the game, the Atari ST version was actually criticized for its visuals by some, because it made the viewable area much smaller.

[6] See Wichman's "A Brief History of Rogue" at http://www.wichman.org/roguehistory.html.

[7] See above note for source.

[8] The full official name is Rogue: The Adventure Game, which is something of a misnomer, as the game had little in common with what most people consider to be an "adventure game," though it does generate new "adventures" each time.

Unfortunately for Epyx and the Rogue team, the commercial versions of the game failed miserably. Wichman blames rampant piracy, but it's likelier that the humble character-set graphics in most versions turned off gamers who were accustomed to advanced sprite-based graphics.

Also, 1983 saw the release of Origin's Ultima III (see book Chapter 23), a game that many considered to be the best role-playing game ever designed for a computer. Its colorful graphics and smooth interface must have made Rogue seem primitive by comparison.

Meanwhile, Daniel Lawrence's Telengard, published by Avalon Hill in 1982 for most computer platforms, offered more-diverse graphics and a similar scheme for procedurally generating the dungeons.

Even Epyx had published a graphical "roguelike" a year earlier: Jeff McCord's Sword of Fargoal for the Commodore 64 and VIC-20. In short, gamers had plenty of CRPGs (computer role playing games) to choose from, including some that offered Rogue-like gameplay, but with more advanced audiovisuals.

However, the real problem seemed to be the "open source" origins of the game. Although it's likely an exaggeration to think of mainframe software programmers as early free software or open source advocates, they did tend to have a much more liberal outlook on sharing and altering each other's code.

Such was certainly the case with Rogue, which had long been available for free in countless public domain versions.

Epyx's Rogue running on an Apple Macintosh SE.

Furthermore, the code had been repeatedly "forked," or taken over by separate development teams and taken in new directions. This growing collection of CRPGs based more or less on Rogue became known as "roguelikes," many of which are still in active development even today.

By the time Epyx got around to publishing the commercial version of Rogue, gamers had several free versions to choose from. The 1986 ports for the Atari ST and Commodore Amiga were even considered embarrassing by some, as critics could point out that more advanced roguelikes such as Hack and Larn were freely available, whereas the commercial Rogue was a typical full-priced game.

Although Rogue wasn't a commercial success, it has still had a considerable impact on the industry, including in its implementation of traditional fantasy conventions. Further, many credit its procedurally generated dungeons as the inspiration for Blizzard's Diablo (see book Chapter 4, "Diablo (1996): The Rogue Goes to Hell").

But we can also find its influence in Japanese imports such as Square's Chocobo's Dungeon 2 (1999) for the Sony PlayStation and Square Enix's Final Fantasy Fables: Chocobo's Dungeon (2008) for Nintendo Wii. However, the sectors where it seems to have the strongest influence is the UNIX and later Linux communities, where Rogue and its derivatives remain a staple of the gaming scene.

Screenshot from Final Fantasy Fables: Chocobo's Dungeon.

Although a comprehensive list of all the roguelikes would likely strain the patience of most readers, it's worthwhile to mention at least a few of the most influential: Moria, Hack, Larn, and Ancient Domains of Mystery. Moria, as the name implies, is themed on J. R. R. Tolkien's literary work, The Lord of the Rings, and debuted in 1983.

The original version was authored by Robert Koeneke, a student at the University of Oklahoma. Unlike the original Rogue, Moria has a persistent town with six shops where players can buy equipment.

Another popular and more recent roguelike named Angband is ultimately derived from this game. Hack, which first debuted in 1982, was authored by Jay Fenlason and three friends. Hack was noted for its clever interaction with the gameworld and its creatures. Slaying and then eating a leprechaun, for instance, will teleport the character to a random location.

This game became the basis for NetHack, a 1987 game that was one of the earliest to utilize the Internet in its development. It was also the basis for Dreamforge's Dungeon Hack, published by SSI in 1993 for the PC. Dungeon Hack integrated Hack's gameplay into the popular Eye of the Beholder engine developed by Westwood Studios.

Noah Morgan's Larn (1986), offered a persistent starting level and a town with a bank, school, shop, and a tax office. Players must pay taxes if they play the game again after winning it (finding a potion to cure the character's daughter); the game's difficulty also increases.

Thomas Biskup's Ancient Domains of Mystery (ADOM) was released in 1994, and is probably the most complex of the lot. It offers quests, skills, and a selection of ten races and twenty characters classes. There are multiple ways to win the game, but the goal is always to stop Chaos from invading the land of Ancardia. There are, of course, many, many other roguelikes, which vary widely in theme and quality.

(click for full size) There are countless "roguelikes," or games that base most of their gameplay concepts on Rogue. Shown here, in various PC versions, are NetHack (top left), Angband (top right), Larn (bottom left), and Ancient Domains of Mystery (bottom right).

There have also been several attempts to recreate Rogue or one of its many derivatives with superior audiovisuals. Two examples of these are Jaakko Tapani Peltonen's NetHack: Falcon's Eye (2002) and Hansjoerg "Hajo" Malthaner's Iso-Angband (2003). Both of these depicted the dungeons in isometric perspective, and also boasted sound effects and music. Unfortunately, neither is in active development now, though Falcon's Eye lives as a fork called Vulture's Eye.

An ambitious and still ongoing project is Scourge, which offers a four-character party and quality audiovisuals.[9] Although some gamers might think these games are huge leaps forward for Rogue, it's also possible to see them negatively.

The task of creating custom graphics for each object and creature in these games is a considerable undertaking that may very well distract developers from what most Rogue fans consider essential: the gameplay.

At least some fans of Rogue may also be resistant to advanced audiovisuals on principle. Malthaner, for instance, felt his project failed because of "acceptance. Not technical issues; these were solvable -- but acceptance was low.

Some people were almost openly hostile towards the idea of a graphical frontend."[10] Purists continue to insist that the essence of Rogue is its gameplay; all efforts to "improve" the audiovisuals merely amount to a distraction, rather like trying to play chess with extremely elaborate pieces.

Falcon's Eye (top) and Iso-Angband (bottom) are two of many efforts to update roguelike gameplay with advanced graphics. The graphical frontends haven't seemed to catch on with many roguelike fans, however, who seem to prefer character-set graphics.

Although Rogue fans may clash over such issues, all agree that the game's main appeal is its gameplay. At its best, Rogue represents an addictive and compelling hack-and-slash type of experience. Unlike the majority of CRPGs, it's an easy game to pick up and play for a few minutes while waiting for a bus.

The character-set versions are also quite easy to get running on even the most limited hardware. The purists may have a good point; the lack of advanced audiovisuals does allow one to better appreciate the more abstract, mathematical nature of the genre. Indeed, perhaps the best way to think about Rogue is as the CRPG (or at least the "dungeon crawler") boiled down to its very essence.

This approach might explain its enduring appeal after so many advances in audiovisual technology, as well as why so many talented programmers continue to explore its potential.

[EDITOR'S NOTE: On sister alt.weblog GameSetWatch, John Harris' long-running @Play column presents perhaps the foremost extended analysis of the Rogue-like genre, in a series of longform articles.]

[9] For information about Scourge, visit http://scourgeweb.org/tiki-index.php.

[10] This quotation is from Malthaner's private e-mail correspondence with the authors.