Going Rogue

07 Jul

When a beleaguered Netscape announced in January of 1998 that it would release the source code to its browser for everyone to tinker with and improve upon, the news shook the worlds of technology and business to their foundations. This open-source “revolution,” as even many in the mainstream press took to calling it, had sprung up seemingly out of nowhere to challenge the conventional wisdom and perhaps the very livelihood of traditional tech giants like Microsoft. For the next several years, you couldn’t open a trade journal or a newspaper’s business section without seeing some mention of the open-source movement and its leading exemplar, the robust and yet totally free — in all senses of the word — operating system Linux. Linux and other software like it was, an eye-opening number of people said, destined to destroy Microsoft’s vaunted Windows monopoly any day now.

The movement’s Little Red Book came in the form of Eric S. Raymond’s 1997 essay “The Cathedral and the Bazaar.” Originally presented as a comparison of a top-down versus a bottom-up methodology in the context of open-source projects, the central metaphor quickly got blurred in the minds of the public into a broader comparison of closed source versus open source, with Raymond’s tacit acquiescence. In this telling, the cathedral was Microsoft’s software-development model, in which a closeted priesthood bestowed programs upon a grateful populace on its own terms and on its own schedule. The bazaar was the hacker way, in which the people came together in a spirit of delightfully chaotic egalitarianism to make software for themselves, sharing their source code in the name of the greater good. “No closed-source developer can match the pool of talent the Linux community can bring to bear on a problem,” wrote Raymond. “The closed-source world cannot win an evolutionary arms race with open-source communities that can put orders of magnitude more skilled time into a problem.” Thanks to Linux and the other open-source tools it enabled, he predicted elsewhere, Microsoft’s eagerly anticipated Windows 2000, the latest incarnation of its server-grade NT operating system, would “be either cancelled or dead on arrival. Either way, it will turn into a horrendous train wreck, the worst strategic disaster in Microsoft’s history.”

Alas, Raymond proved a less effective prophet than pundit. Not only was it not a failure upon its eventual release, but Windows 2000 evolved in 2001 into the consumer-grade Windows XP, by many standards the most successful single version of Windows in history.

Like that of all revolutions that have passed their heyday of strident ideology, the most extreme rhetoric of the late 1990s open-source movement can seem overheated if not downright silly today, the blinkered product of a tiny strata of metaphorical inside cats who have concluded, rather conveniently for themselves, that the most important social-justice campaign of their age is one that can be waged from behind their keyboards and monitors, just the place where they happen to feel most comfortable. As for the ideas they introduced into the public discourse: they were real, valid, and in many ways incredibly valuable, but in the end they would be woven into the fabric of existing corporate-software production practices rather than burning down the old ways wholesale.

For rigid ideology seldom makes a good fit with the real world; pragmatically mixed national economies, for example, succeed vastly better than dogmatically capitalist or communist ones. Similarly, instead of continuing to sort itself into two opposing camps at eternal loggerheads, the modern software ecosystem has learned to take the best from both sides to wind up with a sort of mixed economy of its own. The cleverest actors have learned to combine the cathedral and the bazaar in ways that maximize the strengths of each: Google builds its proprietary Web browser Chrome atop an open-source engine known as Chromium; Apple constructed the OS X desktop on the solid foundation of an open-source operating system known as Darwin; Android mobile phones and tablets have Linux at their core. Even Microsoft now embeds an optional “Linux subsystem” into Windows, as the cats lie down with the dogs.

The reasons for open source’s failure to more comprehensively conquer the world aren’t that hard to divine; they’re actually front and center in some of the movement’s founding principles. The editors of the grandiosely titled 1999 anthology Open Sources: Voices from the Revolution — one of those books whose very name clues you into the window of time in which it was published — wrote that “most open-source projects began with frustration: looking for a tool to do a job and finding none, or finding one that was broken or poorly maintained. Eric Raymond began fetchmail this way; Larry Wall began Perl this way; Linus Torvalds began Linux this way.” The latter two of these projects at least have remained among the most essential of the workhorses that make the Internet function, strong arguments for the superiority of the open-source model for developing some types of software.

But it appears that the same is not true for all types of software. A model in which programmers create only the programs that they most want to have threatens to yield a universe of software which is interesting and attractive only to programmers. Even Eric Raymond had to acknowledge that the production of software with mass appeal is only partially a “technical problem.”

It’s [also] a problem in ergonomic design and interface psychology, and hackers have historically been poor at it. That is, while hackers can be very good at designing interfaces for other hackers, they tend to be poor at modeling the thought processes of the other 95 percent of the population well enough to write interfaces that J. Random End-User and his Aunt Tillie will pay to buy. Computers are tools for human beings. Ultimately, therefore, the challenges of designing hardware and software must come back to designing for human beings — all human beings.

Open source has never entirely made this leap. It’s for this reason that its biggest success stories have come in the realm of back-end software rather than user-facing applications. Witness the long, frustrating history of “Linux on the desktop,” which, in an echo of the old hacker joke about strong artificial intelligence, has been perpetually just a few years away from world domination ever since the late 1990s. There is no theoretical bar to visual designers and experts in ergonomic psychology joining open-source projects, and in some times and places this has even happened. And yet the broad field of open source is still dominated by programmers writing software for themselves and for one another.

Game development joins graphical user interfaces as another notable area where the bazaar model doesn’t quite seem to do the trick. The open-source methodology excels at solving purely technical problems, but the making of a great game is a technical problem only in part — usually, not even the most important part. Consider the case of one of the most critically lauded games of the late 1990s, Valve’s Half-Life. It was a triumph of design and aesthetics, not of technology; its engine was borrowed from id Software’s two-and-a-half-year-old Quake, a technological showstopper in its day which has aged far less gracefully. It would seem that the best way — or perhaps the only way – to create a great game from whole cloth is through a priesthood with a strong and distinctive design and aesthetic vision.

Those open-source games which have become relatively popular have tended to build upon previous game designers’ visions in much the same way that Chrome is built on Chromium: think FreeCiv or Open Transport Tycoon Deluxe, worthy projects that are nevertheless more interested in making workmanlike technical improvements to their inspirations than bold fundamental leaps in design. The open-source movement has had the most pronounced impact on gaming in the form of tools, both for making games and for playing them. I could never have embarked with you on this journey through history that we’ve been on for over a decade now without the likes of DOSBox, ScummVM, UAE, VICE, and many, many other open-source emulators and utilities of all descriptions. I am deeply grateful to the many talented programmers who have given their time to them in order to keep our digital past accessible. Still, they do remain purely technical projects, not creative ones in the sense of the games which they enable to run on modern hardware.

The one ghetto of gaming where open-source projects have been able to forge a strong design and aesthetic sensibility all their own — a sensibility with no obvious antecedents in commercial, closed-source games — turns out upon examination to be not quite the anomaly it might first appear. The “roguelike” sub-genre of the CRPG dates all the way back to 1980, well before the modern open-source movement came to be. But, like that movement, it was a product of an institutional-computing hacker culture that had been around since the 1950s, in which proprietary software was regarded as not so much immoral as simply unheard of. It stands today as a fine example of open source at its best — and equally of what it does less well. Call it the exception that proves the rule.

In Hackers: Heroes of the Computer Revolution, his classic chronicle of the first few decades of institutional hackerdom, Steven Levy writes about the appeal that Adventure, a game that would lend its name to an entire genre, held for the first people to play it on the big multi-user DEC computers of the late 1970s.

In a sense, Adventure was a metaphor for computer programming itself — the deep recesses you explored in the Adventure world were akin to the basic, most obscure levels of the machine that you’d be traveling in when you hacked assembly code. You could get dizzy trying to remember where you were in both activities. Indeed, Adventure proved as addicting as programming…

Rogue, a game which would lend its name to a sub-genre that had even more appeal to the programming mindset, was itself a direct outgrowth of Adventure, with a couple of key elements added to the mix.

Michael Toy and Glenn Wichman were undergraduates at the University of California, Santa Cruz when they first encountered Adventure. Like so many others, they were absolutely entranced. The only drawback was that, once they finally beat the game for the first time, there wasn’t much more to be done; the puzzles were always the same, meaning that beating it again became a rote exercise. And there weren’t yet any other games like it. So, the pair started to talk about creating a game of their own, one that would play a little bit differently. What if, rather than building their game around a collection of pre-crafted set-piece puzzles, they made one that would offer up a new world to the player every single time through the magic of random procedural generation? That way, you could keep playing it forever, even after beating it once or twice or a dozen times. Even Toy and Wichman themselves would be able to have fun with it, given that they too would never know what sort of world they would be entering next.

But what exactly might such a game look like in practice? It wasn’t at all clear; the problem of describing a procedurally generated world in English prose like that used by Adventure was effectively insoluble in the context of the time. Then Toy stumbled upon a new programming library for the Unix operating system (the predecessor to and inspiration of Linux). The brainchild of a University of California, Berkeley student named Ken Arnold, “curses” let you arrange text however you wanted on a terminal screen, letting you change the contents of any one of the 1920 cells that made up a typical 80-character by 24-line display any time you wanted to; this made it possible to reserve different regions of the display for different sorts of information. Earlier games which hadn’t had access to curses, such as Adventure, had had to content themselves with teletype-like interactions: a continuous scrolling stream of text which, once fired at the screen, could only be forgotten. But curses changed all that at a stroke. You could use it to put up menus, maps, charts, and just about anything else you could write or draw using the ASCII character set, updating them all independently of one another.

It gave Toy and Wichman a viable path forward with their fondly imagined infinitely replayable game. For, while textual descriptions of a procedurally generated world were a nonstarter, showing a symbolic, visual representation of one using curses was another matter.

Avid players of tabletop Dungeons & Dragons, Toy and Wichman tried to recreate on the computer the dungeon-delving expeditions they enjoyed with their friends, exploring a network of rooms and tunnels filled with monsters to fight, traps and other hindrances to defuse, and treasures to collect. Whereas the main dish of Adventure had been set-piece puzzles, with only a side dish of dynamic logistical challenges — an expiring light source, an inventory limit, a pesky wandering thief with a sharp sword — the nature of their game meant that it would have to be all logistics. In making this switch, they half-accidentally invented not just the first roguelike but one of the first CRPGs, full stop. We cannot give them complete credit for that genre, mind you: other proto-CRPGs were being created at the same time on the PLATO system at the University of Illinois and on the earliest home microcomputers as well, as other Dungeons & Dragons fanatics also tried to bring the tabletop experience to the computer. Still, by all indications Toy and Wichman made the leap without knowing what anyone else was up to.

It was Wichman who came up with the name of Rogue:

I think the name just came to me. Names needed to be short because you invoked a program by typing its name in a command line. I liked the idea of a rogue. We were coming from a Dungeons & Dragons background, but we were creating a single-player game. You weren’t going down into the dungeon with a party. The idea was that this is a person going off on his or her own. It captured the theme very succinctly.

To depict their world, Toy and Wichman invented the iconography (textography?) that has remained the standard for roguelikes to this day. The walls of rooms were made from horizontal and vertical dashes (“-” and “|”), the tunnels between them from hash marks (“#”), doors from plus signs (“+”), treasure from dollar signs (“$”), monsters of varius types from any and all letters and symbols that weren’t already being used for something else. The focus of it all was your titular rogue, depicted as a forlorn little at-sign (“@”) adrift in this sea of promise and danger. The textual austerity of it all could become weirdly atmospheric. “You’d see a letter ‘T’ on the screen and it would startle you, because you knew it was a troll,” says Wichman.


The goal of the game was to find a MacGuffin called the Amulet of Yendor, hidden 25 dungeon levels or so deep, and return it to the surface. Doing so would require fighting ever more dangerous monsters, building up your character as you did so in classic RPG fashion, both through the experience points you gained from killing them and the equipment you collected. From the first, Rogue was intended to be hard — hard enough to challenge the very people who had made it. This is another quality that has remained a core value of the sub-genre which Rogue invented.

You didn’t know what the stuff you found actually did. Would that yellow potion restore your health, or would it kill you instantly? The safest way to know for sure was to use an “identification” scroll on your new finds, but such things were rare and precious, and ironically had to be themselves identified first. In a pinch, you might just have to try on that new ring or armor and see what happened, praying as you did so that it wasn’t cursed.

Food was the most essential resource of all; while you could eat the corpses of many monsters, some of them would make you sick and some of them would get their posthumous revenge by outright killing you. (Roguelikes are a bit like the old saw about the Australian Outback: everything in them seems to be able to kill you.) The only way to have a chance of winning was to play the game over and over again, slowly ferreting out its secrets and devising optimal strategies in the course of dying again and again and again. Even once you got really good, the difference between success and failure could still come down to sheer dumb luck, as “CRPG Addict” Chet Bolingbroke noted in his articles about the game: “Sometimes you might find a two-handed sword +1 on the first level; other times, you’ll find three poison potions and a cursed dagger.” Rogue‘s own co-creator Glenn Wichman admits that he has never legitimately won it.

Rogue, in other words, flagrantly violated almost all of the modern rules of progressive game design: it was unfair in countless ways and about as unwelcoming to newcomers as a game can be. It was a comedian telling jokes at the poor player’s expense, its later levels stocked with rust monsters that instantly destroyed her hard-won magical armor (until she learned to take it off before fighting them) and rattlesnakes that poisoned her (until she learned that the only practical way to combat them was to chuck whatever junk was to hand at them from a distance). And death was an irrevocable state. Although you could save a game of Rogue and come back to it later, this was intended only for the purpose of resuming an interrupted session: the save file was deleted as soon as you restored it. There were no second chances in Rogue; a single ill-considered move, or a single errant key press, or just a simple stroke of random bad luck, could and usually did erase hours of careful, steady progress.

And yet people found it strangely compelling. This was doubtless partially down to the times; there weren’t a lot of games available to play, which meant that the amount of time and energy required to get good at this one could seem more like an advantage than a disadvantage. But there was also more to it than that, as is indicated by the survival of the roguelike sub-genre right down to the present day, with all of its legendary difficulty intact. Rogue seemed to scratch a different itch than most games, a rash from which hackers seemed particularly prone to suffer. Very few successfully retrieved the Amulet of Yendor, but that only made the prospect of doing so that much more tempting. In the hyper-competitive culture of hackerdom, beating Rogue became a badge of honor almost on a par with writing some super-useful, super-elegant program that made everyone else jealous.

All of this didn’t happen instantly. Like most games on the big institutional computers, Rogue was a work in progress for years after the first version of it went up at UC Santa Cruz, probably in 1980. In 1982, Michael Toy got kicked out of the university for spending too much time tinkering with Rogue and not enough keeping up with his classwork. He took a job in UC Berkeley’s computer lab instead, splintering the partnership that had taken Rogue this far. Wichman now dropped off the scene, to be replaced at Berkeley by, of all people, Ken Arnold, the very hacker whose curses library had inspired the initial creation of Rogue. Toy and Arnold continued to expand and refine the game until they left Berkeley in 1984.

It was during this period that Rogue got really popular, spreading far and wide with the Unix operating system on which it ran, by now the overwhelming hacker favorite. Rogue became an almost equivalent touchstone of hacker culture, being played obsessively everywhere from Bell Labs to the Nevada Test Site. The game’s creators were thrilled when they learned that both Ken Thompson and Dennis Ritchie — living gods among hackers, the creators of Unix itself — were major fans of the game; Ritchie jokingly called it the biggest single waster of CPU cycles in computing history. When Toy attempted to commercialize Rogue in 1984 by releasing an MS-DOS port through the publisher Epyx, he felt justified in advertising it as “the most popular game running on Unix” and “the most popular game on college campuses.”

By the time Rogue hit microcomputers, its partial inspiration Adventure had spawned its own thriving corner of the home-computer-games market, where companies like Infocom sold hundreds of thousands of slickly packaged parser-driven text adventures. But home users proved markedly less receptive to Rogue after its belated arrival. Even after Wichman came back on the scene to help Toy make prettier, semi-graphical versions of the game for the Apple Macintosh, Atari ST, and Commodore Amiga, Rogue didn’t sell all that many copies. Wichman could only conclude that the audience that had made it such a hit on the big computers “wasn’t the audience that was looking for games in software stores.” It was a fair assessment: roguelikes would remain staples of hacker culture, but would never make inroads into the flashier commercial-games market.

Epyx’s Rogue was one of the last artifacts of that company’s original, cerebral “Automated Simulations” identity, appearing the same year that Summer Games and Impossible Mission cemented its new image as a purveyor of slick, audiovisually polished, action-oriented titles. Small wonder that Rogue seemed to get lost in the marketing shuffle.

The Amiga Rogue was a graphical affair, but that didn’t do much for its sales.

In this as in so many other respects, Rogue laid down the template for all of the roguelikes to come as thoroughly as Adventure did for its progeny. But there was one important exception, albeit one external to the game itself: Toy, Wichman, and Arnold didn’t release their source code to the public, clinging to the role of the high priests of a cathedral rather than embracing the bazaar model of software development. “In retrospect, it would have been better to share,” admits Arnold. Yet it isn’t that surprising that they didn’t. Open source had yet to become an ideological movement, even among the hardcore hacker contingent to which Rogue‘s fathers belonged. And they did, after all, have hopes of commercializing the game, even if those hopes ultimately failed to come to complete fruition.

As it was, the lack of source code meant that those who dreamed of building a better Rogue had no choice but to start from scratch. Among the first to do so was a group of boys who hung out together in the computer lab at Lincoln-Sudbury High School in Sudbury, Massachusetts, at the dawn of the 1980s. The school’s single modest DEC PDP-11 minicomputer wasn’t wired to the Internet, but the gang nevertheless encountered Rogue early in its history: in the summer of 1981, when their mentor, a young teacher named Brian Harvey, finagled an invitation for them to go out to Berkeley for a few weeks, to see what life was like in the big leagues of institutional computing. One of the kids who went was named Jay Fenlason. He fell in love with Rogue at first sight, managing to play it for about eight hours by his own estimate during the visit. He returned to Massachusetts determined to make a game just like it. He corralled his buddies into an unlikely game-development team, and over the course of the next year they made Hack, working strictly from their memories of the game they had seen at Berkeley.

That initial version of Hack has been lost, leaving behind only scattered anecdotes. However, all indications are that it wasn’t any remarkable advance over Rogue in itself. What made it important — indeed, what changed everything for the nascent roguelike sub-genre — was the decision Fenlason and his friends made to give away not just their executable but their source code as well.

To celebrate their graduation in 1982, the computer-lab gang packaged up the source code to all of the programs they had written, Hack among them, and sent it to an organization called USENIX, a computing-research nonprofit that maintained a file archive for its members. The source bore a simple notice at the top, saying that anyone who wished to was free to make improvements to the software and distribute them, as long as due credit was given to the original creators as well and as long as they shared the updated source. Having done that, the youngsters who had made Hack went their separate ways, having no idea what the game they had loosed upon the world would someday grow into.

At first, their lack of expectations seemed more than justified; while Rogue went everywhere in hackerdom, Hack went nowhere. Then, in early 1984, a thirty-something Dutch mathematician and programmer named Andries Brouwer, who worked at the Amsterdam research center Mathematisch Centrum, chanced to troll through USENIX’s file archive, looking for interesting software. Just as Don Woods had rescued Will Crowther’s incomplete game of Adventure from oblivion back in 1977, Brouwer now stumbled across Hack and did it the same service. He tightened up the code and the gameplay, and then started adding new features, which he tested on his colleagues at Mathematisch Centrum, most of whom became certifiable Hack addicts. Beginning on December 17, 1984, he uploaded each new version to the Internet as well.

Brouwer added the concept of character classes to the game, introducing six of them; no rogue was to be found among them, but they did include the likes of a tourist and an archeologist, evidence of a quirky sense of humor that would continue to mark the game forevermore. He added shops in the dungeon for buying and selling equipment, and made the dungeon deeper; it now went down 40 levels, the last ten a special region called Hell that demanded magical protection from fire and a teleport spell to even enter. No longer did you find the Amulet of Yendor just lying around somewhere down there in the depths; now you had to defeat a Wizard of Yendor to get your mitts on it. To these big enhancements he added a wealth of smaller details that were likewise destined to remain indelible parts of the game, such as a dog or cat companion to accompany you on your expedition and the ability to write messages on the floor for various purposes.

For years, players of Rogue had been sharing their tips and travails on the Usenet group It was here that Brouwer now announced his new roguelike. The community there pounced upon Hack, which, if not clearly better than Rogue, did have the virtue of being subtly different from a game which most of them had already played to death. The volume of Hack-related traffic grew so extreme that, just one month after Brouwer had uploaded his game for the first time, the group came into being to accommodate it. “Please stop posting articles about Hack to and use this new group instead,” wrote a Usenet administrator pointedly.

Brouwer kept his fire hose of additions and improvements spurting until July of 1985, when he pronounced himself satisfied with the game and moved on to other things. But, thanks to the fact that he had honored the wishes of Jay Fenlason and company and publicly released his source code, Hack could continue to morph and grow after his departure in a way that Rogue had not been able to after Michael Toy and Ken Arnold left Berkeley. Ports and modified versions were soon popping up everywhere. It was exciting in a way, but it became a bit too much like the babble of a bazaar. Three hackers, by the names of Mike Stephenson, Izchak Miller, and Janet Walz, decided that a little bureaucracy wouldn’t be amiss. They decided to create a sort of curated version of the game, incorporating changes from anyone who wished to contribute to the project, as long as they were well-coded, worthwhile, and not game-breaking. Because their home base was, they named their version of the game NetHack. Its first official release came in July of 1987; its most recent one as of this writing came out in February of 2023. I suspect that there will be many, many more before NetHack‘s full history can be written.

NetHack is an answer for every player of traditional adventure games who has ever asked why she can’t just bash a door open instead of searching hither and yon for the key.

The semi-anonymous wizards behind the NetHack curtain are known simply as the DevTeam. For 36 years, this rotating cast of characters has maintained and added to the game, making it one of if not the most systemically complex ever created, even as it retains in its canonical version an entirely textual display focused around a little wandering at-sign. Experienced players delight in ferreting out the emergent possibilities provided by the sheer depth of NetHack‘s systems. “The DevTeam thinks of everything,” goes a saying among players.

To wit: use a pair of gloves to pick up a dead cockatrice, a creature which turns any living thing it touches to stone, then bash your enemies with it to turn them to stone. (This technique is known among the NetHack cognoscenti as “wielding the rubber chicken.”) Of course, you’ll need to use a pick axe afterward to separate the statues of your enemies that are left behind from the loot they were carrying…

Or combine a Wand of Polymorph with a Ring of Polymorph Control to eliminate the middleman, as it were, turning yourself into a cockatrice. You can lay eggs in this form, which you can pick up and carry around once you revert to your natural form, throwing them at your enemies like grenades while you gleefully sing “Rainy Day Women #12 and 35.”

The possibilities are endless. NetHack even keeps track of the phases of the moon in the real world and uses them to influence your luck; this leads to devotees clearing their calendars once per month in order to maximize their chances when the moon is full.

NetHack has become an institution of old-school hacker culture, and with it an icon of the open-source movement. None other than Eric Raymond was the first to create an optional graphical skin for the game (a move that prompted considerable controversy). And well before he wrote The Cathedral and the Bazaar, he wrote the first manual for NetHack. Small wonder that it joined Rogue and Adventure as one of the very few games memorialized in 1996’s New Hacker’s Dictionary — edited by, you guessed it, Eric S. Raymond. DevTeam founding member Mike Stephenson has no doubts about NetHack‘s importance, not only as a standalone game but as a model for software development: “We predated open source [as a movement], but I do think we helped to promote the idea of making software available for public use without cost. I think the other thing that really contributed to the concept of open source is that NetHack has, and still does, accept bug reports and feature ideas from anyone.”

NetHack became the standard bearer of the roguelike sub-genre almost from the moment of its first release, and has never had its status in this regard seriously challenged. That said, hundreds of other roguelikes were made after it, and some even before it. The most important among them are arguably Moria and Angband. The former arrived at a complete form already in 1983, when it became the first game of this type to offer an above-ground town to serve as a base for your dungeon expeditions; this gave it a significantly different feel, more like, to put things in the terms of Dungeons & Dragons, an ongoing campaign than a single adventure module. Moria directly inspired 1990’s Angband, a much more complex implementation of the same approach, which, like NetHack, is still in active development today. Some players prefer NetHack‘s relentlessly escalating challenge, others Angband‘s somewhat more relaxed pacing and more free-form structure — but make no mistake, Angband too will kill you in a heartbeat if you let your guard down. And in it as well, dead is dead, permanently.

This roguelike “family tree” shows how the most historically and currently popular games in the sub-genre relate to one another.

This brings us back around to a statement I made at the outset: that roguelikes are the exception that proves the rule of open-source game development — and just possibly of open-source software development in general. The cast of thousands who contribute to them do so in order to make exactly the games that they want to play, which in the abstract is the best of all possible reasons to make a game. The experience they end up with is, unsurprisingly, much like high-wire programming at its most advanced, presenting players with an immense, multi-faceted system to be explored and mastered. And there is absolutely nothing wrong with this.

Still, it does seem to me that roguelikes tend to bring out some of the worst as well as the best of the hacker ethic, what with their insistence that they’re only for the “hardcore” and their lack of empathy for the newcomer. Few things in this world are less attractive than a nerd beating his chest. Robert Koeneke, the creator of Moria, admits that while he was working on it, “if anyone managed to win, I immediately found out how, and ‘enhanced’ the game to make it harder.” Likewise, for every cool interaction to be discovered in NetHack, there’s a cheap, heartless death in store, like stumbling down a staircase whilst carrying a cockatrice and turning yourself to stone, or missing a stirrup whilst trying to mount a horse and breaking your neck, or incinerating yourself by firing off your Wand of Lightning too close to a wall, or getting killed by your own pet dog when you attempt to use your Ring of Conflict to get that nearby band of orcs fighting one another. NetHack is the sort of game that likes to give you a fake Amulet of Yendor, than laugh at you when you scurry all the way back to the surface with it and think you’re about to win.

As with so much in life, one’s relationship to roguelikes comes down to questions of priorities. As someone who likes to play a variety of games, I’ve never done more than dabble in these ones. For the time required to get even minimally competent at them is more than I’m willing to invest in any single game — or that I can invest, if I want to keep doing what I do on this site.

Meanwhile the amount of time and effort required to get good at a game like NetHack is staggering, even if you’re far smarter and more diligent than I am. It took Chet Bolingbroke 262 hours of trying to win at NetHack for the first time — and that was playing in a fashion that many purists would consider illegitimate, by looking up spoilers on the game’s many interconnected components rather than learning strictly through experience, not to mention playing an old version that is much less complex than the current ones. Was it worth the time investment? He has his doubts. “Permadeath just sucks,” he concludes. Even Eric Raymond feels today that NetHack may have gone too far: “There was a natural tendency for the devs to see the game from the point of view of someone who played it constantly and obsessively. Thus, over time, their notion of not making it ‘too easy’ gradually ratcheted up the difficulty level to the point where you really couldn’t enjoy it casually anymore.” NetHack displays, in other words, open-source software’s usual Achilles heel, its developers’ inability to put themselves in the shoes of people who aren’t just like them.

Then again, it isn’t as if this represents some deep moral failing; there’s nothing wrong with being niche. Many or most lovers of NetHack and other roguelikes have never won them and quite probably never will, finding satisfaction merely in the trying, in hoping to get a little further than last time and walk away with some entertaining stories to share. Far be it from me to begrudge them their pleasures. Although I doubt that I will ever become a big fan of roguelikes, I do derive a quiet sort of satisfaction from knowing that things so implacably committed to being their own idiosyncratic selves exist in this world.

And if roguelikes will never go mainstream, that doesn’t mean they haven’t influenced the mainstream. Next time, we’ll learn how one of the most popular of all the slick commercial games of the late 1990s grew out of this odd little corner of hackerdom…

Did you enjoy this article? If so, please think about pitching in to help me make many more like it. You can pledge any amount you like.

(Sources: I highly recommend David L. Craddock’s book Dungeon H@acks: How NetHack, Angband, and Other Roguelikes Changed the Course of Video Games, a treasure trove of information that I have only touched upon here. The CRPG Addict blog is full of stories about what it’s like to actually play Rogue, Hack, 1987-vintage NetHack, 1989-vintage NetHackMoria, and Angband among other roguelikes, along with some more historical notes. I’m immensely indebted to David for all of his original research and to Chet for spending the hundreds of hours on these games that I couldn’t spare.

Other print sources include the books Hackers: Heroes of the Computer Revolution by Steven Levy, The Cathedral and the Bazaar: Musings on Linux and Open Source by an Accidental Revolutionary by Eric S. Raymond, and Open Sources: Voices from the Revolution edited by Chris DiBona, Sam Ockman, and Mark Stone; Byte of March 1984 and February 1987; Acorn User of February 1997; Computer Power User of March 2008. Other online sources include Glenn Wichman’s “Brief History of Rogue,” “The Best Game Ever” by Wagner James Au at Salon, “Playing the Open Source Game” by Shawn Hargreaves, “Freeing an Old Game” by Ben Asselstine at Free Software Magazine, and a retrospective on NetHack by Dave “Fargo” Kosak of GameSpy.

Much more information about all of the games mentioned in this article, and roguelikes in general, can be found at RogueBasin, as can download links for all of them.)


Tags: , , ,

80 Responses to Going Rogue

  1. Andrew Plotkin

    July 7, 2023 at 4:10 pm

    > Whereas the main dish of Adventure had been set-piece puzzles, with only a side dish of dynamic logistical challenges — an expiring light source, an inventory limit, a pesky wandering thief with a sharp sword

    Now you’re just trolling me. :)

  2. PlayHistory

    July 7, 2023 at 4:17 pm

    This was an interesting place to put this story! It is often easy to forget how much of our modern digital ecosystem rests on top of open source. A great shame of mine is to never be a good enough programmer to contribute to these projects which have given me so much.

    Roguelikes and the community around them are both deep fascinations of mine, having lived to see the rise of the roguelike genre in recent years. I really enjoy both the hardcore roguelikes and the modern interpretations, probably stemming from my endless reloading of maps in Diablo II.

    I believe that variants of this game do represent the most plausible path to building “systemic” games in a larger context. Hades showed that you can very effectively build non-linear storytelling from a roguelike base. Hell, even Starfield is adopting the idea of populating pre-built areas procedurally as roguelikes already do.

    I think there’s so much to learn from this genre in terms of engaging player storytelling, despite the bevies of complexities to both development and player entry. The more modern games take from roguelikes, the closer they get to my ideal of a game that truly defies past narrative conventions. That may be an unpopular belief, but I think that’s the way.

    My favorite roguelike is Cataclysm: Dark Days Ahead. Would be interested to hear what others people here enjoy.

    • Dave Rolsky

      July 8, 2023 at 12:01 am

      I have played C:DDA for so many many hours. I even made a CLI tool to download dev releases of C:DDA and launch them on Linux ( Note that it’s probably broken because I haven’t played the game for a while (I tend to play in spurts every few years).

  3. Andrew Plotkin

    July 7, 2023 at 4:29 pm

    > The one ghetto of gaming where open-source projects have been able to forge a strong design and aesthetic sensibility all their own…

    I’m going to split a hair and say that parser IF shares this distinction. Adventure was passed around, picked up by a new developer, ported — it was part of the open-source culture of its time, even though (like Nethack) it precedes the label. The MIT version of Dungeon was treated the same after Infocom took themselves off to chase sales.

    It’s fair to say that Nethack was the big example of an open-source project that stayed at the center of its genre. (Parser IF went commercial, then shareware, then free, but never adopted open-source licenses as a default for game releases.) But if you want an open-source origin *defining its genre’s sensibility*, that’s Adventure in spades.

    • Jimmy Maher

      July 7, 2023 at 8:07 pm

      I think the waters here are too thoroughly muddied by Infocom. Adventure was first, but the IF Renaissance of the 1990s was really focused on the Infocom aesthetic: “interactive fiction” with a degree of literary ambition to go along with the puzzles and neat environments to explore. I don’t really see this as Adventure’s aesthetic at all; whatever literary qualities it has — and it does have a certain minimalist beauty about it — strike me as somewhat accidental. It’s perhaps telling that a lot of people in the mid-1990s were still trying to sell text adventures via the shareware model, trying to recreate the Infocom distribution model in some form or fashion. This was not the case with roguelikes; it just wasn’t done in those communities.

      Put another way: you can recount the history of classical roguelikes without mentioning commercial software at all. You can’t do that in the case of IF. Or, put still another way: the Epyx Rogue was an oddity rather than the game changer and re-focuser (for lack of a real word) that Zork I was.

      And then IF has retained a very book-like authorship model, with one or at most a few authors prominently credited. Roguelikes are far more anonymous, community-driven efforts.

      IF *tools* are a different story, of course…

      • dissolved

        July 8, 2023 at 10:25 am

        But wouldn’t “classical IF” in this context be the mainframe games that birthed commercial games? You could still recount the history of mainframe Zork or the Phoenix games without mentioning their respective commercial iterations, and you could, say, draw a line between mainframe Zork and MUDs while ignoring Infocom entirely.

        • Jimmy Maher

          July 9, 2023 at 7:07 am

          Well, sure, you can draw lines anywhere you want to in the end. But my point is that you can write a pretty solid, comprehensive history of Berlin-interpretation ( roguelikes without mentioning commercial software at all. The IF history you describe, on the other hand, would strike just about everyone who knows anything about the genre as idiosyncratic and incomplete. (Which doesn’t mean it would be bad…)

          • Anthk

            October 18, 2023 at 12:03 am

            Also, well, Inform6 tools (and 7, too) are under a libre license.

            And there’s Spiritwrak if you want to play a Zorkian game, among lots of free as in freedom text adventures such as All Thing Devours, Beyond The Titanic and Supernova.

            OFC most roguelikes are developed under the libre hacker culture and most IF games are under the writer culture, but nowadays IF writers should learn that having your game under a libre license would make far easier to translate it into different languages such as Spanish .

            And, yes, lots of people would love to play Spider and Web or Anchorhead in their native languages.

  4. Ugh

    July 7, 2023 at 5:29 pm

    I’ve been reading for a long time and wow is this ever not what I came here for. “Social justice”? I get plenty of holier-than-thou capitalism shoved at me by Elon already, thanks.

    Typed on a device with a glued-in battery and a proprietary OS. Still read via “the internet,” but Meta and Twitter are finishing that off rapidly.

    What a disappointment.

    • IJMC

      July 7, 2023 at 8:13 pm

      I’m not quite sure what @Ugh above was getting at — but if it was a sense of disappointment in the article’s tone toward open-source software (to say nothing of ignoring the key distinctions between OSS and free/libre software), then I for one share it.

      • Ugh

        July 8, 2023 at 12:51 am

        That was it.

      • Jason Hutchens

        July 10, 2023 at 7:52 am

        Me three. Was a length anti-open-source diatribe really the best lead-in to a history of rogue? Especially when it’s first mentioned over halfway through the article to make the point that rogue was not open-source, but probably should have been?

        Then again, I suppose that the article is itself an open-source work, and suffers from similar failings, so perhaps this was all a clever, multi-layered joke.

    • Brent

      July 8, 2023 at 11:39 am

      As a layman I very much appreciated Jimmy’s well-reasoned explanation. As a counter argument you’ve offered nothing but an extremely off putting, holier-than-thou dismissal of the entire article, ironically proving Jimmy’s point about open-source’s defenders being unable to bring the majority of people on blowers.

      • xxx

        July 9, 2023 at 8:02 pm

        Amen, Brent. As a software developer who’s been part of various open-source communities since the 1990s, I entirely agree with Jimmy’s impression of the open-source “revolution”. It never became what we hoped and didn’t realize that open source approaches were only good for solving certain kinds of problems rather than a universal panacea for fixing software development.

        While the concept of open source is a beautiful thing, it’s always been one of open-source’s great tragedies that so many of its most vocal proponents are awful people. A few examples: operating system creators Linus Torvalds and Theo de Raadt, who belittled and mocked anyone who disagreed with them or failed to appear smart enough; open-source advocate Eric Raymond, a misogynist, homophobic, racist, libertarian wingnut; and GNU founder Richard Stallman, a man completely disconnected from reality who publicly defends pedophilia. (None of this is libel; it’s all a matter of public record.)

        I think that open source succeeded more in spite of these people than because of them. The traits that make someone an effective leader of an open-source project are entirely orthogonal to the traits that make someone a non-awful human being.

  5. Vladimir Kazanov

    July 7, 2023 at 5:38 pm

    Probably worth mentioning that about 10-12 years ago a Renaissance of roguelikes happened. I remember how Brogue, yet another clone of the original game, became a cult of its own. I wasted numerous hours on it! and then indie game scene started looking for inspiration in elements of the genre: Spelunky, Darkest dungeon, Hades to name a few top sellers.

  6. Ido Yehieli

    July 7, 2023 at 6:14 pm

    A bunch of minor points/nitpicks:
    * You once misspelled Wichman’s name as “Bechman”.
    * 80×25 is DOS/Windows standard terminal size, on UNIX the standard size is 80×24 (based on DEC’s VT100 terminal).
    * Among classic traditional roguelikes one of the most dominant is *Dungeon Crawl Stone Soup* (better known by its acronym DCSS), whose design-focus is perhaps more “clerical” than NetHack and Angband’s.
    * In the meanwhile the name “Roguelike” has penetrated the mainstream (see for example the sheer amount of games tagged as such on steam:, albeit the meaning has incrementally drifted with time.
    * A very interesting relative/offshoot which started a sub-genre of its own is *Dwarf Fortress*.

    • PlayHistory

      July 7, 2023 at 6:31 pm

      I have to imagine Jimmy will at least touch upon the particularity of the “roguelike” term in the following Diablo post.

      • Paul

        July 9, 2023 at 6:22 am

        That was my thought immediately upon reading this: “roguelikes would remain staples of hacker culture, but would never make inroads into the flashier commercial-games market.”

        Friggin’ Diablo is basically just a geaphically pleasing roguelike w/o permadeath. Roguelite? That the term? Similar games such as Children of Morta have come out over the years and enjoyed commercial success. To say the genre has completely missed the mainstream is beside the point.

        • Jimmy Maher

          July 9, 2023 at 7:01 am

          Cultural taxonomy is always difficult, but I hew in this article to the Berlin Interpretation of a roguelike: Diablo scores no more than 50 percent on this metric, qualifying it for “heavily influenced by” but not *of* the roguelike sub-genre. (There are actually some pretty huge differences beyond the lack of perma-death that are every bit as notable as the similarities: while roguelikes are known for their difficulty and their hardcore nature, Diablo emphasizes frictionless ease of play above all else. I’d almost go so far as to argue that it’s basically a casual game dressed up in hardcore-gamer clothing.)

          I do suspect the term “roguelike” has gotten especially badly diluted in the context of Steam. That said, I probably should have made these things clearer in the article itself. The nature of this site means that I sometimes assume an old-school perspective by default. Hopefully I’ll have time tomorrow to add a well-considered sentence or two.

          • Jack Brounstein

            July 9, 2023 at 8:23 pm

            This blog, with its video game history focus, might be an exception, but in general I don’t like the term “roguelite” or the insistence that a “roguelike” game has to be a turn-based, tile-based dungeon crawl. Orders of magnitude more people play Spelunky, The Binding of Isaac, and Hades than NetHack, ADOM, or Dungeon Crawl Stone Soup, and it feels wrong to insist that the latter get exclusive use of the term. (None of this is a knock on old-school/classic roguelikes! I’ve played and enjoyed a lot of NetHack. But it’s worth acknowledging that it’s a very niche genre.)

    • Jimmy Maher

      July 7, 2023 at 8:12 pm

      Thanks! Corrections made on the first two points. The others are just a bit out of scope for what I wanted to cover here…

      • Ido Yehieli

        July 8, 2023 at 10:20 am

        Very fair! I meant these more as addendum/comments than corrections :)

  7. Anony Maus

    July 7, 2023 at 6:32 pm

    “Open Source” failed to do any good because it was specifically designed not to. Raymond and O’Reilly specifically crafted the narrative of it as an engineering concept rather than one of moral or ethical policy, in contrast to the earlier Free Software movement.

    Everything after that was entirely predictable.

  8. Jonathan Badger

    July 7, 2023 at 7:19 pm

    I realize that it is a minor point compared to what you are actually talking about, but it is incorrect to say “In his telling, the cathedral was Microsoft’s software-development model, in which a closeted priesthood bestowed programs upon a grateful populace on its own terms and on its own schedule”. Both the cathedral and the bazaar were referring to open-source projects and weren’t talking about closed software from Microsoft and the like.

    The difference was projects in the bazaar model (like Linux) accepted contributions from the general public (if the maintainers agreed that the contribution was good), where the “cathedral” projects were things like GCC and Emacs, where the code was developed by a small team, and while they released their code, they weren’t very open to outside contributions.

    • Jimmy Maher

      July 7, 2023 at 8:23 pm

      Good catch! In my defense, I don’t think that Raymond himself wasn’t always clear on the distinction after he became known as the spokesman for open source. Edit made. Thanks!

    • Pedro Timoteo

      July 8, 2023 at 6:32 pm

      Yeah, I was going to stay the same. At least in the context of the late 90s, “the cathedral” typically referred to open source programs that were developed by a relatively small team of “priests” who, while they released the code as a tar.gz, developed their software in a closed (almost mysterious) way, with relatively sporadic releases (sometimes with years between them), instead of the most current source being available as it’s being developed. In other words, if the latest version is two years old, that’s the most recent code available to the public.

      Another classic example (though this may have changed since the 90s, I haven’t kept up) is mentioned in this very post: NetHack.

  9. UIUCAlum

    July 7, 2023 at 7:26 pm

    This is a neat and informative article. Minor correction: You seem to have confused the University of Chicago with the University of Illinois at Urbana-Champaign when discussing the PLATO system.

    • Jimmy Maher

      July 7, 2023 at 8:25 pm


  10. stepped pyramids

    July 7, 2023 at 7:32 pm

    For what it’s worth, the newer generation of roguelike fans and devs (and there’s a great deal of crossover there still) are not totally dedicated to “hardcore” game design. Smaller and/or easier and/or fairer takes on the genre are all over the place these days. And even a difficult game like Dungeon Crawl Stone Soup is winnable by an average player; I think I got my first win in a month or so. (I’ve never won NetHack legitimately and I’ve been playing it for two decades.)

  11. Nate

    July 7, 2023 at 9:08 pm

    Eric Raymond is a very problematic individual who has said racist and batshit things. I don’t think he’s qualified to speak for open source software and was always only a self-nominated spokesperson.

    Typo: “grahical”

    • Jimmy Maher

      July 8, 2023 at 7:28 am


  12. Michael

    July 7, 2023 at 9:24 pm

    “Game development joins grahical user interfaces”


  13. Keith Palmer

    July 7, 2023 at 10:50 pm

    Epyx’s Rogue was available on the Tandy Color Computer 3, so I did experience it relatively early on. I have to admit, though, to figuring out that if I put a write-protect tag on the diskette at just the right moment, I wouldn’t be affected by “permadeath…” (So far as the game not selling well as a commercial product went, I also have to admit to the impression my family’s copy just might have been handed along from my uncle…) I fear my final admissions might have to be that my dabbling in more complicated “roguelikes” haven’t gone that far, and that I was a bit more inclined to nod along at this piece’s comments on “open source” than some of the earlier commentators.

  14. Andrew McCarthy

    July 7, 2023 at 11:23 pm

    Great article, though I must confess I’ve never played any roguelikes, having no interest in the genre whatsoever. I’ve experienced enough arbitrary cruelty in video games from vintage Sierra and (I have to say it) Infocom adventures to last a lifetime, thank you. ;)

    The respective difference in genre terminology when referring to the progeny of Adventure and Rogue always amuses me. I like to joke that in a parallel universe, instead of “adventure games” and “roguelikes”, we’d be playing “rogue games” and “adventurelikes”.

    The “cathedral” metaphor for Windows proved unusually apt in one respect: the “XP” of Windows XP bears a resemblance to the Christogram “Chi-Rho” symbol (☧) of early Roman and later Byzantine Christianity, made up of the first two letters of Christ’s name in Greek, Chi (Χ) and Rho (Ρ). And of course the code name for XP while in development was “Windows Cairo”. Perhaps somebody at Microsoft was having a laugh about it…

    • Vince

      July 31, 2023 at 5:20 am

      There is a fundamental difference between the “cruelty” in adventure games and roguelikes, though.

      In the former, if you are stuck at a particular unfair puzzle, the game literally stops happening. You might put additional hours in the game trying to get unstuck, but you won’t become “better” at the game, you’ll just increase the chanches of solving that particular puzzle, while having a thoroughly miserable time.

      In (good) roguelikes every failed run slowly builds towards a winning one, and there is an instrinsic satisfaction in seeing your progress over time. Bosses that once seemed unbeatable are now easy, levels that seemed unreachable are now familiar.

      Modern roguelikes/lites are maybe not part of the “mainstream” but definitely they managed to carve a big niche for themselves as some of thf modern games offering an actual challenge.

  15. Brett Coulstock

    July 7, 2023 at 11:31 pm

    A thoughtful article that really puts IF and Roguelikes into perspective as reflections of a culture and philosophy and time and place. I can imagine some people bristling a little at this, but I think you’ve made your points and justified them expertly.

    Another spelling error: “… combat them was to chunk whatever junk was to hand … ”

    Should have been “chuck”?

    • Jimmy Maher

      July 8, 2023 at 7:32 am

      In the American South, it’s common to say “chunk.” But I’m not sure how much of the Texas country boy is left in me after all these years away, and “chuck” is more classy. ;) Thanks!

  16. Alex Freeman

    July 7, 2023 at 11:38 pm

    I would add that, generally speaking, the open-source approach seems better suited for programs under a certain size and complexity. Once you get to operating systems, it becomes a real trade off. Linux and Windows have certain advantages over each other with neither consistently better. MacOS is probably the best designed, building on the open-source BSD OS, but it lacks the software support of Windows.

  17. Dave Rolsky

    July 8, 2023 at 12:08 am

    Besides Cataclysm:Dark Days Ahead, which PlayHistory noted, my other favorite old school rogue-like game is definitely ADOM (Ancient Domains of Mystery). I probably played hundreds of hours of this around 20 years ago.

    It’s a really interesting game with an overworld, multiple dungeons, multiple victory conditions (some are insanely complex), many classes _and_ races, and very deep systems (you can farm stuff, dip things in potions, etc. which I think comes from NetHack). It also has some non-random areas mixed in, like specific levels in dungeons or towns. But I’ll admit I save-scummed because I just don’t have the patience to play the first few hours of the game a hundred times over.

    There’s a graphical version on Steam these days which I haven’t tried.

  18. Martin

    July 8, 2023 at 12:41 am

    So where do games like Temple of Apshai intersect with these games, or are they just two separate things? Apshai came out in 1979, I think.

    • Jimmy Maher

      July 8, 2023 at 7:41 am

      Apshai does bear a lot of surface similarities to Rogue in terms of presentation and interface, as do the early Dungeons & Dragons-style games on PLATO. But it’s pretty clearly a case of parallel development rather than influence. There are some major differences when it comes to Apshai in particular: pre-crafted rather than procedurally generated dungeons, and a much stronger interest in storytelling, to the point that Apshai came with a book of paragraphs to be read at the appropriate times. I wouldn’t place Apshai in the roguelike category.

  19. John

    July 8, 2023 at 2:34 am

    At some point circa 1989 I played Rogue on a friend’s Amiga for perhaps five to ten minutes. It was, I’m sorry to say, the least compelling Amiga game he ever showed me. I’m older and wiser now of course, so I’d like to think that if I knew then what I know now I’d be able to look beyond the graphics and appreciate the gameplay. Alas, I have my doubts. That ten minutes remains to this day the sum total of my old-school Rogue-like experience. I’m not too good for text-based games. I spent a lot of time on MUDs when I was in college. Nor am I too good for Rogue-likes or Rogue-lites. In the last week I’ve played both Crypt of the NecroDancer and Streets of Rogue. But old school Rogue-likes have a certain reputation for difficulty for difficulty’s sake and for complexity for complexity’s sake that I find off-putting. I think, looking back, that I could have been very into Rogue or some other Rogue-like if there had ever been a period where it was the only game I could get my hands on. The reason I got so into MUDs in college is that I didn’t own a computer or a video game console. I could, however, to the computer lab, sit down at a terminal, and Telnet to my favorite MUD. If Rogue or Nethack had been available on the school’s Vax I might have gotten absorbed in one of them instead.

  20. gabriel

    July 8, 2023 at 3:04 am

    almost all your guesses to history are wrong. nothing of it was natural or organic. most companies openly spent millions on lawyers and culture pushes to convince a new generation that just making money out of someone’s browser engine was ethical… and also to convince the authors of that browser engine that licensing on non-gpl was better for humanity and themselves because having some big corporation that don’t do evil using your project will bring fame and fortune to you and professional contributors to your project… heh.

    the fact you gladly parrot their very false arguments from the time as some historical truth is a testament to how effective is to control the narrative.

  21. Nate

    July 8, 2023 at 5:06 am

    Roguelikes had a second spurt of growth as Unix shell accounts for Internet access (and then Linux on PCs) proliferated in the 1990s.

    A few minor nits:

    No comma needed here
    “, and made the dungeon deeper”

    It took me a long time to realize “chunk” meant “throw” here. Since “to hand” is a verb in other contexts (i.e. to pass an object), maybe the simpler “throw” would be clearer?
    “only practical way to combat them was to chunk whatever junk was to hand at them”

  22. Nate

    July 8, 2023 at 5:19 am

    Oh and also it was interesting to me that “archeologist” is a less preferred spelling in most English-speaking countries.

  23. Sean Curtin

    July 8, 2023 at 6:21 am

    I think the word you were looking for might have been “cloistered” rather than “closeted.”

    • Jimmy Maher

      July 8, 2023 at 7:51 am

      While either will work in a pinch, I think “closeted” is the better choice, actually. “Closeted” means “functioning in secret,” while “cloistered” means “secluded from the world.”

  24. Jaina

    July 8, 2023 at 7:10 am

    Michael Toy and Glenn Wichman were undergraduates at the University of California, Santa Cruz when they encountered Adventure for the time. Like so many others, they were absolutely entranced. ”

    For the first time?

    • Jimmy Maher

      July 8, 2023 at 7:53 am


  25. Laertes

    July 8, 2023 at 10:12 am

    One of the most epic sessions I’ve ever had on any computer game whatsoever was in Dungeon crawl stone soup some years ago. After several days and plenty of hours I was killed in the swamp by a large slime creature. The death was totally my fault as I disregarded all caution believing my character was almost invincible. I found out the hard way that was not the case. I should have known better by then.

    Anyway, I ended sweating and with my heart beating as if I had just run a marathon. And that in a turn game in which you can think each movement for as long as you please.
    I still keep the character file, it was and still is my best game ever:

    Dungeon Crawl Stone Soup version 0.10.3-5-gf0cc9df character file.

    144242 Tpit2s XI the Executioner (level 18, -5/162 HPs)
    Began as a Hill Orc Berserker on Sept 25, 2012.
    Was the Champion of Trog.
    Slain by a large slime creature (14 damage)
    … on Level 5 of the Swamp on Oct 5, 2012.
    The game lasted 11:41:48 (43357 turns).

    Roguelikes are really something special.

  26. J

    July 8, 2023 at 2:05 pm

    > roguelikes would remain staples of hacker culture, but would never make inroads into the flashier commercial-games market.

    In the US and Europe maybe, but Japan has plenty of proper (by whatever definition) roguelikes on consoles that did middling-to-well. Koichi Nakamura, founder of Chunsoft and creator of the Dragon Quest franchise, was a big Rogue fan leading to the Mystery Dungeon series. The first game sold well and and spawned multiple subfranchises continuing to this day. There are also several unrelated roguelikes that did relatively well like Azure Dreams.

    • Sniffnoy

      July 9, 2023 at 5:15 pm

      Oh, yeah, this is important; they weren’t just commercial, they were on consoles even! And eventually these started getting localized for the US as well, even if they were never nearly as big here.

  27. Robert

    July 8, 2023 at 3:01 pm

    I highly enjoyed the article, but I’m not sure you’ve proven your central tenant vis-a-vis Open Source being related to the success of roguelikes.

    * The massive success of of Rogue and its clones/next generations pre-dated the publication of any code.

    * The roguelike “family tree” include major entries such as ADOM and Omega which either never released their code or released it very late.

    * Hack/Nethack/etc… are very successful due to the “rotating cast of characters” who are updating the code, but I’m not sure that it wouldn’t be just as successful (or rather, still successful but in a different way) with a different licensing model which was less open.

    So, just going by the material in this article, I don’t see why Roguelikes are particular examples or exemplars of anything open-source-y.

    • stepped pyramids

      July 9, 2023 at 7:52 pm

      Angband is the more effective example here (and NetHack vs. Angband represent their own cathedral vs. bazaar comparison). Its design is very friendly to the development of variants and cross-pollination of ideas between said variants.

      I think without Angband and NetHack you wouldn’t have a concept of a “roguelike”, just a handful of games that cited this old game Rogue as an influence. I’m also skeptical that closed-source roguelikes like ADOM and Dwarf Fortress (adventure mode) would have caught on without an existing community of roguelike fans.

  28. Bog

    July 8, 2023 at 3:31 pm

    “Google builds its proprietary Web browser Chrome atop an open-source engine known as Chromium; Apple constructed the OS X desktop on the solid foundation of an open-source operating system known as Darwin”

    I’m hopelessly unqualified to speak to either of these points, but this part reads as if Chromium and Darwin were initially developed independently of Google and Apple.

    Chromium is the open-source codebase for Chrome, but the engine is the open-source Blink, which was forked from Apple’s open sourced WebKit engine, which is a fork of KDE.

    Likewise, the note about Darwin seems to suggest that it was an independent project that Apple took off of the shelf and built OS X on top of, but Darwin was the platform developed by Apple specifically for OS X and based on NeXTSTEP, so the real open-source precursors would be BSD and Mach.

    I’m afraid it’s a bit misleading to present the relationship between Chrome and Chromium as Google making workman like improvements on an open-source base. The truth is much less charitable: Chrome’s market share a Google’s involvement in Chromium basically ensures that Google has an extremely outsized role in web standards. W3C or any other standards body can do what they want, but if Google doesn’t include a feature (or even just includes it, but disabled and hidden in a settings panel), then it’s DOA.

    Apple almost certainly open-sourced WebKit and Darwin with a similar intent (although given their market share at the time both were developed, it was probably more of a survival strategy; Jobs named Darwin, so genetic fitness is seemed to be a concern for Apple), but once their portable devices started printing money and eating market share, open source pretensions to Darwin basically evaporated.

  29. Bogdanov

    July 8, 2023 at 3:51 pm

    “For rigid ideology seldom makes a good fit with the real world; pragmatically mixed national economies, for example, succeed vastly better than dogmatically capitalist or communist ones.”
    “there’s nothing wrong with being niche”

    These two sentences shall be painted in every parliamentary chamber :D

  30. Rowan Lipkovits

    July 9, 2023 at 1:35 am

    Realising you were going to be discussing Rogue I was primed to expect mention of BSD and felt a little robbed when you mentioned that OSX was built on Darwin and not looking a little further back to its NeXTSTEP or BSD ancestors, as Bog notes. Also while pointing out how popular Rogue was among Unix users I felt mention might have been warranted that given the Berkeley employment arrangement of its author, Rogue was actually included with BSD distributions for five years starting with v4.2 in 1982. Having source code available is one way to be “spread far and wide” among Unix users, but having the game actually be a pack-in title sure can’t hurt!

    The Half-Life example, while correct, feels a little odd in context given id Software’s historical commitment to eventually opensourcing their engines, but granted you are making a different point there.

    The list of excellent open source game tools seems incomplete without acknowledging MAME, but I appreciate that the list could go on for pages.

    Reference to The CRPG Addict’s experience of Rogue also seems incomplete without noting (ok, it’s worth a footnote at most) that his experience of playing that specific old game through to completion was a primary factor in the establishment of his long and rigorous project of CRPG scholarship.

    “why she can’t just bash a door open” Hey, that looks familiar! I submitted that screenshot to MobyGames 15 years ago. You’re all very welcome!

    Making ESR such a central part of your article is certainly a way to lure out wildly polarized reader responses in 2023! We’re just lucky that RMS didn’t have a role to play in this particular story…

  31. Glorkvorn

    July 9, 2023 at 9:14 am

    NetHack seems like a “you had to be there” experience. I could see it being fun when played cooperatively in a computer lab with a group of friends: sharing knowledge, commiserating over deaths, and laughing together at the dumb jokes late at night. It doesn’t feel fun at all when you try to play it now by yourself, as an adult. Either you go into it blind, in which case it’s just an endless series of inexplicable deaths, or you look up some spoilers/walkthroughs on the internet and you lose all the process of exploration. Also, yes, it’s just too complex and bloated now. Learning it now feels like… well, it feels like trying to learn Linux administration for the first time, and getting up to speed on code that was once simple and elegant but has since been growing nonstop for decades.

    Rogue I think still holds up today. It’s relatively short, simple, and self-contained. It’s hard, but at least it fights fair- you can pretty much discover everything by playing normally, you don’t have to read the developer’s minds. And a lot of the challenge is just randomness. On the one hand, even with perfect play you’re still probably going to die. On the other hand, you can play impulsively and still make it pretty far just by luck. It works surprisingly well as a casual mobile game that you can play to relax without putting in much effort. Just roll with the punches and keep restarting until everything goes your way. It also helps that Rogue doesn’t try to cram in too much of the wacky 80s nerd humor like NetHack does.

  32. Ronald Record

    July 9, 2023 at 3:15 pm

    When evaluating the success or failure of Open Source software I think it is important to realize that most of the phones in the world run on open source software, much of the Internet is based on open source software, most of the backend servers and cloud systems are based on open source software, and Linux is everywhere embedded in devices of all kinds.

    I would assert that proprietary software has failed to capture the majority of markets and is now a niche endeavor by the few large corporations who adhere to the false belief that secrecy has value.

    Other than maybe a difference of view on the success of open source software, I found this article to be extremely well written, of great interest, and quite informative. Thanks!

  33. Jack Brounstein

    July 9, 2023 at 8:15 pm

    Without getting too far afield of the blog’s topics, but the line about “a tiny strata of metaphorical inside cats who have concluded, rather conveniently for themselves, that the most important social-justice campaign of their age is one that can be waged from behind their keyboards and monitors, just the place where they happen to feel most comfortable,” feels incredibly relevant to the people today, on all sides of the spectrum, who seem to seem to view getting into arguments online as the most important political act.

    The last paragraph is of course setting up the story of beloved Sega Genesis game ToeJam & Earl, right? I joke, but I played the game as a child, many years before experiencing NetHack and learning more about classic roguelikes, and was surprised when I revisited it how much of the genre it kept despite the graphics and real-time gameplay. (Also, it was designed by Greg Johnson, who previously worked on Starflight.)

  34. Jack Brounstein

    July 9, 2023 at 8:26 pm

    Sorry, one more thought: Does setting up Diablo mean that the blog’s about to move into coverage of 1997? I was still looking forward to the Battlecruiser 3000 AD and Her Interactive articles (although looking now, I guess Her Interactive’s Nancy Drew series didn’t start until 1998).

    • Jimmy Maher

      July 10, 2023 at 8:05 am

      Diablo sits splayed neatly on the border between 1996 and 1997. Although I believe its official release date was January 3, 1997, there are many credible reports of it being available in stores in some regions at least between Christmas and New Years. I think I’m going to consider it a 1996 game, mainly because it will fit nicely with the two articles that precede it in an eventual ebook. But it will mark the end of 1996 coverage. I should have a “state of the blog” post with ebooks, etc., within a month or so.

      I’m still planning to write on Her Interactive; in fact, I was at the Strong Museum in Rochester in January, and returned with a treasure grove of documents. But I decided to push that to 1998, to tell the story from the beginning until the first Nancy Drew game, the point at which the company really found its footing after a lot of flailing around, in one article.

      I’m afraid I did decide to pass on Battlecruiser in the end. It’s an amusing story, but not much more than that in the bigger context of gaming, and I’m afraid the final article would just become an exercise in making fun of the game and its rather, shall we say, *eccentric* maker. I think it’s better to focus on more positive topics. (But if I ever put together a rogue’s gallery of tacky 1990s game advertisements — something I’ve mused about doing from time to time — *that* ad will definitely be included.)

  35. Tom Dolezal

    July 10, 2023 at 4:52 pm

    Hi Jimmy… near the top of this article, under the graphic of Epyx’ Rogue game box cover, “Automated Systems” is referenced. The actual name of the company pre-Epyx was “Automated Simulations”.

    See the first sentence in this entry:

    I think I had most of Automated Simulations’ games for the Apple II early on. Temple of Apshai and Hellfire Warrior were favorites that burned up way too much of what should have been more college study time.

    • Jimmy Maher

      July 10, 2023 at 5:41 pm

      Good catch. Thanks!

  36. DDG Ahab

    July 11, 2023 at 12:59 pm

    I’ve understood that 1982 was when Fenlason began developing Hack, and that he released it to USENIX around 1984. Nethack Wiki says as much.

    The extant DOS port “hack121” probably stays close to Fenlason’s design, though it’s impossible to know for sure. The main differences from Rogue are addition of stores, being able to ascend to previously explored levels (whose state persists, unlike Moria where levels are regenerated each time), and just having more monster types and more types of treasures. This takes off a lot of the time pressure that Rogue’s hunger meter forces on you, but also slows the pace down and makes death costs you a lot more time.

    • Jimmy Maher

      July 11, 2023 at 1:45 pm

      I’m going to defer to David Craddock’s book, which correlates the submission of Hack to USENIX with some pretty significant life events of the principle players: graduation from high school, and also the departure of Brian Harvey, the computer-lab gang’s mentor, from his teaching position at Lincoln-Sudbury. But I did just learn that Jay Fenlason works with an acquaintance of mine — small world, huh? — so I’ll try to confirm, and update the article if necessary.

    • Adam Sampson

      July 12, 2023 at 12:25 am

      The earliest Brouwer sources for hack published on Usenet on 17th December 1984 include Jay Fenlason’s Original_READ_ME file, which starts: “This is export hack, my first semester programming project.”

  37. Adam Sampson

    July 12, 2023 at 12:41 am

    “1996’s New Hacker’s Dictionary — compiled by, you guessed it, Eric S. Raymond”

    It might be better to say “edited by” or “revised by”, given that much of the content is from the Jargon File, originally compiled by a large cast of characters at MIT, Stanford, and other research centres – many of whom you’ve already mentioned, e.g. Don Woods. Guy Steele was the primary editor of the first print edition in 1983, and Eric Raymond’s 1991 and 1996 editions were substantial revisions of that. The 1991 edition already mentions nethack.

    • Jimmy Maher

      July 12, 2023 at 12:23 pm

      Fair enough. Thanks!

  38. Doug Orleans

    July 12, 2023 at 2:13 am

    It’s a pleasant surprise to see Brian Harvey’s name pop up here. He was a beloved lecturer at UC Berkeley and taught the intro CS class for several decades. I did an independent project for him when I was an undergrad circa 1991, and I also got into the PC version of Hack (and later nethack) around that time, but I never knew about his involvement (however indirect). As part of that intro CS class I learned to use JOVE, which was a stripped-down version of Emacs written by another of his former high school students. I continued to use it for several years after college until eventually moving up to full Emacs. That same CS intro class also had an extended exercise that involved building a small text adventure game in Scheme, based on a similar curriculum from MIT, but modified to use Berkeley landmarks and in-jokes.

  39. Alianora La Canta

    July 14, 2023 at 10:40 pm

    I’m very happy to see Rogue get mentioned; having gone well past the point where it was originally released, I’d resigned myself to thinking it was never going to be covered by the blog. There is a certain amount of nostalgia from remembering playing it quite a bit when I was a child (even though I never got past level 10). I still enjoy Nethack (my record there is level 18), but it is notably more difficult because of the combinatorial explosion.

    Nethack is in the Museum of Modern Art, resulting on DevTeam member Jean-Christophe Collet providing one of my favourite quotes in software design, “You should always write clean code that you won’t be embarrassed by, 35 years later, when it ends up in a museum.”

    It also reminded me of when I saw a review for Nethack in the British edition of PC Gaming World in March 2000, presented as if it was a new game, complete with a technical specification box including “3D Graphics: Don’t be silly”. For anyone wondering, the only other game in that issue that got five stars was The Sims by Maxis. Now that’s a diametrically opposite approach to game development philosophy…

  40. Anthony Noel

    July 30, 2023 at 3:48 pm

    Once more poor Beneath Apple Manor is unmentioned in the history of the rouge like. It does most of what rouge did and did it two years earlier and as a commercial release.

    • Busca

      July 30, 2023 at 7:18 pm

      If you haven’t already read it, maybe check out the 2012 coverage of BAM by the CRPGAddict who mentions the similarities with Rogue and had an exchange with creator Don Worth – with the latter also showing up in the comments.

    • Jimmy Maher

      July 31, 2023 at 12:45 pm

      While its similarities to later roguelikes are certainly interesting, there is no sign that Beneath Apple Manor directly influenced them. This makes it a bit hard to shoehorn into an article like this one, which is concerned specifically with the roguelike tradition in institutional computing. A similar complaint to yours could be made on behalf of the early PLATO CRPGs, which also bore some distinct similarities to Rogue; one always has to draw a line somewhere in order to craft readable articles. That said, you might want to have a look at David Craddock’s book Dungeon Hacks, which devotes its whole first chapter to Beneath Apple Manor, based upon a lengthy interview with its creator.

  41. Carl Grace

    October 19, 2023 at 3:32 am

    A bit late to the game, but the Nevada National Security Site was named the Nevada Test Site until 2010 so the name “Nevada Test Site” makes more sense in your statement.

    • Jimmy Maher

      October 20, 2023 at 12:21 pm


  42. Nifft

    December 16, 2023 at 11:50 am

    I don’t uderstand the conclusion of this article, regarding the open-source debate. Looking at the IT world nowadays, in the 2020’s, it is clear that the advocates of open-source were completely right. Even Microsoft has done a complete u-turn.

    • SpookyFM

      March 17, 2024 at 12:30 pm

      I’m a bit late to the party but I kind of agree with you. If I understand Jimmy correctly, his conclusion isn’t really that open source would never be successful but that it’s unlikely to produce a polished product suitable for a broad user base rather than for hackers. I would argue that projects such as Firefox, Thunderbird, Open/LibreOffice, Ubuntu (as well as countless other Linux distributions, all put together from myriad _other_ open source projects), MediaWiki (the software that Wikipedia and many other Wikis are running on), the Signal messenger and many others show that there are too many exceptions to sensibly call that a rule.

      (As some anecdotal evidence: Sure, I’m a, let’s say, semi-nerd and thus closer to the target group that Jimmy argues is more likely to use and benefit from open source projects. But my parents’ – decidedly non-nerds – home computer has been running on Ubuntu Linux for I think a decade now, they’re doing all the updates etc. themselves, and I can count on two hands the times that I had to help with some OS-related problem. I don’t think that number would’ve been significantly lower if I had installed Windows or gotten them a Mac. And I put the computer together for around 300-400 Euros at the time and I haven’t had to do a hardware upgrade since, and all of the software they’re using is and was free. That’s hard to achieve with a commercial solution.)

      What I will concede is that all of these projects have one thing in common: They are all mainly funded and in many cases managed by a corporation or a well-endowed foundation, moving them closer to a cathedral situation. So, no, open source hasn’t been a wildfire revolution but the model has been successful. And I expect it to become even more so: The EU for example has finally woken up to the idea that all aspects of our lives are increasingly depending on digital services, and they are starting to force actors in the space to make their infrastructures more open and interoperable, which will necessitates _at least_ APIs to be open and documented. Some towns and cities here in Germany are committing to using either exclusively open source software or at least making their own data and procedures open and comprehensible to citizens. The activists pushing for these changes are often people who have been on the forefront of the open source movement (e.g. the Electronic Frontier Foundation in the USA or the Chaos Computer Club here in Germany). So it’s been a slow burn, an endothermic reaction rather than a revolution, but I don’t think you can sensibly call it niche.


Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.