§46   A short history of interactive fiction

The history of interactive fiction in the twentieth century has yet to be written. One outline might be as follows: an age of precursors and university games, 1972–81; the commercial boom, 1982–6; a period of nostalgia among Internet users for text while the industry completed the move to graphical games, 1987–91; and the age of the Usenet newsgroup rec.arts.int-fiction and its annual competition, of shorter stories moving away from genres and puzzles, 1992–9.

§46.1   Precursors and university games 1972–81

Perhaps the first adventurer was a mulatto slave named Stephen Bishop, born about 1820: “slight, graceful, and very handsome”; a “quick, daring, enthusiastic” guide to the Mammoth Cave in the Kentucky karst. The story of the Cave is a curious microcosm of American history. Its discovery is a matter of legend dating back to the 1790s; it is said that a hunter, John Houchin, pursued a wounded bear to a large pit near the Green River and stumbled upon the entrance. The entrance was thick with bats and by the War of 1812 was intensively mined for guano, dissolved into nitrate vats to make saltpetre for gunpowder. After the war prices fell, but the Cave became a minor side-show when a desiccated Indian mummy was found nearby, sitting upright in a stone coffin, surrounded by talismans. In 1815, Fawn Hoof, as she was nicknamed after one of the charms, was taken away by a circus, drawing crowds across America (a tour rather reminiscent of Don McLean's song “The Legend of Andrew McCrew”). She ended up in the Smithsonian but by the 1820s the Cave was being called one of the wonders of the world, largely due to her posthumous efforts.

By the early nineteenth century European caves were big tourist attractions, but hardly anyone visited the Mammoth, “wonder of the world” or not. Nor was it then especially large, as the name was a leftover from the miners, who boasted of mammoth yields of guano. In 1838, Stephen Bishop's owner bought up the Cave. Stephen, as (being a slave) he was invariably called, was by any standards a remarkable man: self-educated in Latin and Greek, he became famous as the “chief ruler” of his underground realm. He explored and named much of the layout in his spare time, doubling the known map in a year. The distinctive flavour of the Cave's names – half homespun American, half classical – started with Stephen: the River Styx, the Snowball Room, Little Bat Avenue, the Giant Dome. Stephen found strange blind fish, snakes, silent crickets, the remains of cave bears (savage, playful creatures, five feet long and four high, which became extinct at the end of the last Ice Age), centuries-old Indian gypsum workings and ever more cave. His 1842 map, drafted entirely from memory, was still in use forty years later.

After a brief period of misguided philanthropy in which the caves were used as a sanatorium for tuberculosis patients, tourism took over. By the twentieth century nearby caves were being hotly seized and legal title endlessly challenged. The neighbouring chain, across Houchins Valley in the Flint Ridge, opened the Great Onyx Cave in 1912. By the 1920s, the Kentucky Cave Wars were in full swing. Rival owners diverted tourists with fake policemen, employed stooges to heckle each other's guided tours, burned down ticket huts, put out libellous and forged advertisements. Cave exploration became so dangerous and secretive that finally in 1941 the U.S. Government stepped in, made much of the area a National Park and effectively banned caving. The gold rush of tourists was, in any case, waning.

Convinced that the Mammoth and Flint Ridge caves were all linked in a huge chain, of perhaps four hundred miles in extent, explorers tried secret entrances for years, eventually winning official backing. Throughout the 1960s all connections from Flint Ridge – difficult and water-filled tunnels – ended frustratingly in chokes of boulders. A “reed-thin” physicist, Patricia Crowther, made the breakthrough in 1972 when she got through the Tight Spot and found a muddy passage: it was a hidden way into the Mammoth Cave.

Under the terms of his owner's will, Stephen Bishop was freed in 1856, at which time the cave boasted 226 avenues, 47 domes, 23 pits and 8 waterfalls. He died a year later, before he could buy his wife and son, and achieve his ambition of farming in Argentina. In the 1970s, Crowther's muddy passage was found on his map.

One of Pat Crowther's caving companions was her husband, Will, who had already used computer plotters to draw the group's maps. He takes up the story:

I had been involved in a non-computer role-playing game called Dungeons and Dragons at the time [c. 1975], and also I had been actively exploring in caves … Suddenly, I got involved in a divorce, and that left me a bit pulled apart in various ways. In particular I was missing my kids. Also the caving had stopped, because that had become awkward, so I decided I would fool around and write a program that was a re-creation in fantasy of my caving, and also would be a game for the kids … My idea was that it would be a computer game that would not be intimidating to non-computer people, and that was one of the reasons why I made it so that the player directs the game with natural language input, instead of more standardized commands.

(Quoted in Dale Peterson, Genesis II: Creation and Recreation with Computers, 1983.) It's hard not to feel a certain sadness that the first adventure game is shaped by these two lost souls, Bishop and Crowther, each like Orpheus unable to draw his wife out of the underworld.

Crowther's program (c. 1975), then, was a simulation of the Bedquilt Cave area, owing its turn-based conversational style to a medieval-fantasy adaptation of tabletop wargaming: E. Gary Gygax and Dave Arneson's Dungeons and Dragons (1973–4). Nor was the program without precedent, either in computing – ‘Hunt the Wumpus’ (Gregory Yob, 1972) was a textual maze game, while ‘SHRDLU’ (Terry Winograd, 1972) had a recognisably adventure-like parser – or in literature, where OuLiPo and other ludic literary genres, especially in France, had tried almost every permutation to make physical books more open-ended: Raymond Queneau's Cent mille milliards de poèmes (1962) cut its pages into strips so that the lines of ten sonnets could be mingled to form 1014 different outcomes.1 But the OuLiPo writers, and earlier futurists, had thought more in terms of clockwork than the computer: the literature machine's unashamed mindlessness a provocation to the reader, in whom associations will be triggered. Italo Calvino (in his 1969 lecture Cybernetics and Ghosts):

It will be the shock that occurs only if the writing machine is surrounded by the hidden ghosts of the individual and of his society.

To all intents and purposes, then, ‘Advent’ had invented a new category of computer program and of literature. The aim was to explore, with five treasures hidden below and only a few of the more “natural” puzzles as obstacles, such as the snake, the dwarves and pirate, the first of the mazes and the limited battery span of the caver's essential companion, the carbide lamp. Like the real Bedquilt, the simulation has a map on about four levels of depth and is rich in geological detail:


There are photographs of this chamber and of the column that descends to it, which is of travertine, an orange mineral found in wet limestone. The game's language is loaded with references to caving, to “domes” and “crawls”. A “slab room”, for instance, is a very old cave whose roof has begun to break away into sharp flakes which litter the floor in a crazy heap.

Working at SAIL, the Stanford Artificial Intelligence Laboratory, in the spring of 1976, Don Woods discovered Crowther's game among a number available to be played across the burgeoning (110-computer) network ARPANET, the child of a shotgun wedding in 1969 between university and Department of Defense PDP-10 (and some other) computers. The PDP-10, whose character set did not include lower case letters – hence the capitals above, although elsewhere in the book quotations from ‘Advent’ have been normalised – was widely found to be a “friendly” computer for recreational use, but more to the point it was a time-sharing computer on which individual users could run programs much larger and more complex than traditional games like the PDP-1's ‘Spacewar!’. With Crowther's eventual blessing, but working entirely independently, Woods reworked the caves and stocked them with magical items and puzzles, liberally ignoring the original style from time to time. Much of the game's classic quality comes from the tension between the original simulation, the earnestly discovered caves with their mysterious etched markings and spectacular chambers, and the cartoonish additions – the troll bridge, the giant's house, the Oriental Room, the active volcano. Crowther contributed an austere, Tolkienesque feel, in which magic is scarce, and a well-judged geography, especially around the edges of the map: the outside forests and gullies, the early rubble-strewn caves, the Orange River Rock. Some of Woods's additions, such as the bear, were sympathetic but others, such as the vending machine for fresh lamp batteries, clashed against the original. But their strange collaboration is somehow consistent. Stretching a point, you could say that there is a Crowther and a Woods in every designer, the one intent on recreating an experienced world, the other with a really neat puzzle which ought to fit somewhere.

By 1977 tapes of ‘Advent’ were being circulated widely, by the Digital user group DECUS, amongst others, taking over lunchtimes and weekends wherever they went. The idea spread, and diffused, as it surprised members of the general public who were shown it by friends. In Tracy Kidder's Pulitzer prize-winning book The Soul of a New Machine (1981), a journalist's-eye-view of the building of a new model of company-sized computer, ‘Advent’ appears as an addiction, but more: while the engineers use the program as a convenient endurance-test, for Kidder it is a cypher for an absorbing inner world and, perhaps, an emerging personality. Another fascinated visitor, the television producer Patrick Dowling, created The Adventure Game (BBC1 and BBC2, May 1980 to February 1986): by a curious coincidence, his first choice as puzzle-deviser and scriptwriter was Douglas Adams, then at the BBC but as it turned out unavailable. Adams will reappear later. In the Game, Earth-people were tested by the alien Argonds by being made to explore rooms stocked with items and quite difficult puzzles in hope of finding drognas (the currency of Arg), payment of which might placate His Highness the Rangdo, who had adopted the body of an aspidistra plant with a tendency to shake and roar when irritable. (And there may have been other anagrams of “dragon”.) A recurring puzzle was a simple adventure game running on, naturally, a BBC Micro, so that each week viewers would see fresh contestants sit at the keyboard and twig, eventually, that the “scarlet fish” was a red herring.

Most of the university departments then connected to the ARPANET specialised in computer science, where any program is an invitation to develop a further one. At Essex University, England, Roy Trubshaw and Richard Bartle developed the concept of ‘Advent’ into ‘Essex MUD’ (which ran from late 1979 to September 1987 and continues in different forms even today). MUD was a Multi-User Dungeon to which remote users logged on during the night, competing sometimes unkindly with each other – killing another player netted 1/24th of their points, which must otherwise be earned by the troublesome business of finding treasures and dropping them into a swamp – to become “wizards” in a fantasy landscape anglicised slightly by the presence of a thatched cottage. To early phone-line networks such as British Telecom's Prestel Gold and CompuServe, running MUDs was (briefly) lucrative, and in a sense the Internet-connected “deathmatch” tournaments of today's games like ‘Quake’ are the legacy of MUDs.

In 1979–81, “game assemblers” were written in at least three departments to make new “Adventure-like programs” – the plural “adventures” seems not yet to have been used. Chris Gray and Alan Covington's “Six/Fant” at the University of Alberta, Canada, and the UCLA Computer Club's “Dungeon Definition Language” (which later evolved into Tim Brengle and Ross Cunniff's ADL (1987)) deserve mention. At Cambridge University, England, however, the assembler by David Seal and Jonathan Thackray may have been the first “adventure design system” to be used more widely than by its creators. Here is some typical code, allowing the player to jump up to a hole only if carrying the chair and in the room which actually has the hole:


The assembler was used to build sixteen games which were the chief recreation on “Phoenix”, the IBM 370 mainframe used by undergraduates and academic staff throughout the 1980s. These were large, computationally expensive games, traditional in form and very difficult, played outside of prime time when research palled. “Well go and do some work then” is the parting shot of ‘Fyleet’ (Jonathan Partington, 1985). Titles tended to be distinctive one-word commands, supposedly the names of ancient lands. Some games were later released by Acornsoft and some later again by Topologika, so that these are sometimes vaguely called “the Topologika games”. But to anyone who was there, they are as redolent of late nights in the User Area as the soapy taste of Nestlé's vending machine chocolate or floppy, rapidly-yellowing line printer paper. Adam Atkinson (author of ‘Nidus’, 1987), who still has faint sketch-maps drawn on that paper more than ten years ago, has recently worked with Paul David Doherty and Gunther Schmidl to recover much of the Phoenix source code; many of the games have now returned to play through mechanical translation to Inform.

‘Advent’ had no direct sequel as such, but for the five years to 1982 almost every game created was another ‘Advent’. The standard prologue – middle game – end game form would have, for prologue, a tranquil outside world (almost always with a little building offering two out of three of a bunch of keys, a bottle and a lamp); the middle game would be a matter of collecting treasures from a cave and depositing them somewhere, while the end would be called a “Master Game”. The secret canyons, cold spring streams, wizards' houses, passive dragons, bears, trolls on bridges, volcanos, mazes, silver bars, magic rings, lamps with limited battery power, octagonal caverns with exits in all directions and so forth recur endlessly in a potent, immediately recognisable blend. Publicity surrounding the notorious Ace Paperbacks pirate edition (1965) of The Lord of the Rings had helped make Tolkien's epic an American campus classic of the late 1960s: ten years later, most of the cave games can be seen to have superficial Tolkienisms, with elves, dwarves (note the spelling) and dungeons called Moria. Unsurprisingly, then, the first book adaptation in interactive fiction seems to have been ‘Lord’ (Olli J. Paavola, c. 1980), initially a mainframe game at Helsinki.2 (The earliest ARPANET connections outside America were to Britain and Scandinavia.) ‘Lord’ took pains to be faithful to the text, even to including the ballads.

You are standing now in Longbottom where Tobold Hornblower once lived, the one who first grew the true pipe-weed in his gardens, about the year 1070 according to Shire reckoning. To the south-east is a narrow path.

It was characteristic of Tolkien, who died in 1973, to think the Hobbit tobacco industry as “real” as the story of the Ring, but it has been characteristic of Tolkien's imitators and adaptors to prefer orcs and magic: ‘Lord’, though never completed, has a true authenticity. Even here, though, the myth of the cave game – the underground labyrinth linking the computers – is as strong as the myth of Middle Earth. ‘Lord’ also has a Flathead coin and a postage stamp, in clear reference to ‘Dungeon’, the 1978 mainframe distribution of the game before and afterwards called ‘Zork’.

At one extreme of the cave game is ‘Adventureland’ (Scott Adams, 1978), the first commercial game to reach the home: a tiny set-piece for cassette tape-based microcomputers, written under vicious memory constraints. “I'm in a temple” is as detailed as it gets, but Adams's games are distinguished by weirdly errant grammar, a wide vocabulary and a talent for arranging diverse objects in a room to portray it:

I'm in a dismal swamp.
Obvious exits: North, South, East, West, Up.
I can also see: cypress tree – evil smelling mud – swamp gas – floating patch of oily slime – chiggers

At the other extreme is ‘Acheton’ (David Seal, Jonathan Thackray, Jonathan Partington, 1978–80), probably the largest game in the world in 1980, with 162 objects in 403 locations. (The title is a confection of Acheron, the underworld, and Achates, a character from the Aeneid.) Here is the lodestone room:

You are in a large featureless room whose walls are composed entirely of a black magnetic material. Your compass seems incapable of fixing on any direction as being north. Several passages lead off to other parts of the cave.

This might easily be a room from ‘Advent’: and for all that they vastly differ in scale, ‘Acheton’ and ‘Adventureland’ are recognisably the same game.

As ‘Advent’ spread through universities, so it was often reworked and altered. As with Chaucer's Canterbury Tales, the vast number of mutated versions is evidence of popularity not just with the audience (players) but with those who told the tale (programmers). Chaucer's original manuscript is lost but all 83 surviving variants are thought to derive from a single version copied from it. Here, it is Crowther's 5-treasure original (c. 1975) which is gone, and all known forms of ‘Advent’ build on Woods's 15-treasure extension of June 1977, further diluting Crowther's contribution, that is, the simulation aspect of play. (For a “filiation” almost as complicated as Chaucer's, that is, a family tree showing how the many versions relate to each other, see ftp.gmd.de.) Most of these extensions are inferior works, making nervous and minor additions, but three deserve passing mention. Don Woods made a further extension in Autumn 1978 to a “20-treasure version (Revision 2)”, which he still considers definitive: it made the modest addition of a reservoir and cliff, scoring from 430 rather than 350 points. David Platt's 550-point version (1979) has a “Valley of the Stone Faces” and a puzzle bringing the volcano into play. Like Platt, David Long (1978) also felt the need to add a sword-in-the-stone puzzle: Long's 501-point version has some painful incongruities, such as a Wumpus and a telephone box, but is actually not too bad.

A port by Jay Jaeger for a (substantially souped-up) kit-built Altair 8800 is claimed to be the first microcomputer version: if so, it was not alone for long. Microsoft and Apple, unequal titans of the future but contenders even then, followed with ‘Microsoft Adventure’ (Gordon Letwin, 1979) and ‘Apple Adventure’ (Peter Schmuckal and Leonard Barshack, 1980) for Apple II and TRS-80. Early commercial versions were faithful or trimmed, although Microsoft added a “Software Den”, north of the Soft Room, containing computers and a bearded programmer whose “spells help keep this cave together”. (Cf. the RAM location in ‘Adventureland’, or the appearance in person of the programmers of ‘Enchanter’.) Level 9's multi-platform ‘Colossal Adventure’ (1983) also has a classic feel but makes a confident extension, with a fleshed-out landscape above ground including a spire and a hawthorn wood, and a more satisfying end-game. The authors of the little-known but rather good Spectrum 128K version ‘The Serf's Tale’ (Nigel Brooks and Said Hassan, Players, 1986) seem familiar with Level 9's, and add a mild intrigue (in cut-scenes, the player searches a dead body for keys, and is helped by an innkeeper) on the way to the caves. ‘The Serf's Tale’ takes embellishment to baroque extremes: “You are in a splendid chamber shaped like the inside of an Arabian tower. The walls are frozen rivers of orange stone that curve gently up to a shadowy apex some thirty feet above your head. From this a huge stalactite hangs like an inverted spire above the centre of the room.” One wonders if Crowther would still have known his place.

§46.2   The commercial boom 1982–6

“Then Adventure hit MIT and everything changed.” The response of a disparate group of students, an improvised imitation called ‘Zork’, led to the founding of Infocom, in June 1979, which at its height six years later employed a hundred people: its mainframe, “a fleet of red refridgerators” (Brian Moriarty), had the electricity bill you would expect “if you were running an aluminum smelter” (Marc Blank). An engaging image, but extensive testing and packaging were also critical in establishing Infocom as a quality brand in a self-created niche market. Infocom's glory years have been romanticised by talk of its free soda, aloha shirts and the Tuesday lunchtime meetings of the Implementors of games – of whom there were never more than ten, and who were by no means as free to do whatever they liked as their image suggested. Too little credit has been given to department heads who were at least as responsible for Infocom's artistic texture, notably John Prince (book editor, lunch host and low-key manager of the Implementors) and Liz Cyr-Jones (chief of testing, and the only woman to substantially influence the creative process). “The staff dresses casually, and it appears as if some of them have slept in their clothes, if they have slept at all…”, wrote Richard Dyer in the Boston Globe (6 May 1984), the most perceptive and least wide-eyed of their many journalist visitors. There was pain on the way, particularly in the discovery that games were to be the only viable product of the former MIT Dynamic Modelling Group, and not merely an interim line. Unsavoury corporate dealings after a buyout by another ailing company, together with the Implementors' own faltering belief in text as the medium and the exhaustion of key members of staff, made the winding-down of 1988–9 unnecessarily dispiriting. But former hands mostly look back on the heyday as a happy, one-time thing, like a summer romance.

Infocom was dominant for a period in the higher-end, chiefly American market: in 1985 it always occupied several of the top twenty positions in the SoftSel Hot List – industry-wide sales charts run by a major US distributor – and one game held the number one slot for nine months. But the company was not nearly so visible outside the USA, where disc drives were less affordable, and in any case had no monopoly on the basic idea. “The ‘adventure boom’ is on – witness the rash of new programs, books and even a specialist magazine.”3 Although many were short-lived in what was something of a cottage industry, Hans Persson's catalogue cites 329 production companies. For instance, in the UK, Acornsoft made an early start: based in Cambridge and with close links to the university, it had a ready-made supply of adventure designs and quickly released reworkings of the Phoenix mainframe games for the BBC Micro – Acorn's computer, but built to spearhead a national computer literacy campaign supported by television programmes. Some of Acornsoft's titles made tidy business (one of its authors earned royalties of around 35,000 on games originally written with no thought of profit), but in the end a market limited to a single model of microcomputer was insufficient to support a large games company.

Looking back at the early microcomputers is like looking at the fossils in ancient shale, before evolution took out three quarters of the species, some of them weirder than anything living today. The market had been entered by Apple, Commodore, Tandy (1977), Atari, Exidy (1978), Acorn (1979), Sinclair (1980), Osborne, IBM (1981) and a dozen others, whose machines were mutually incompatible in that software could not easily be transferred from one model to another. Text adventures were an exception, using little of the more complex hardware (for graphics and sound) which really differentiated designs: also because a typical adventure program is 90% map, text and other data tables, so that only 10% would need rewriting to move to a new machine. The more specialised design companies got this division down to a fine art, and ‘Zork I’ was offered for 23 different microcomputers.

Infocom was incomparably the largest of these specialist houses, the others directly employing five or six people at most. On paper the main rivals were, in America, Scott Adams's Adventure International (seventeen games, 1979–85); in Australia, Melbourne House (seventeen highly variable games from different sources); in Britain, Level 9 (twenty games 1981–91, founded by the brothers Pete, Mike and Nick Austin) and later Magnetic Scrolls (seven games 1984–90, founded by Anita Sinclair and Ken Gordon, 1984). In practice their markets – geographical and by computer ownership – overlapped little so that competition came not from each other, but from different genres of game.

Right through the “golden age of text adventures”, around two-thirds of the perhaps 1,000 adventure games published mixed graphics with text, usually in the form of a picture accompanying each room description. Sierra, still a major player in the games industry today for its long-running King's Quest series, began this trend as early as 1980 with ‘Mystery House’ (Roberta Williams). Sales were so good that board game companies (Avalon Hill, Games Workshop) dabbled in the market, and spin-offs from books, film and television began to appear, though play seldom really engaged with the subject matter. When ‘Dallas Quest’ (James Garon, 1984) confronts the player with Miss Ellen, a bugle and a rifle, it is easy to guess which of the three will prove to be decoration. Spinnaker/Telarium Software's adaptations of two classics of science fiction, Arthur C. Clarke's ‘Rendezvous with Rama’ (Ronald Martinez, 1984) and Ray Bradbury's ‘Fahrenheit 451’ (Len Neufeld and Byron Preiss, 1984) deserve mention, though, as does the same company's ‘Amazon’ (1984), by the novelist and screen-writer Michael Crichton, later to become famous for Jurassic Park but already hot property in Hollywood. Thomas M. Disch, another novelist of real powers, went through a wild surge of enthusiasm writing ‘Amnesia’ (1986), to be followed by total disillusion when it was not marketed and received as a novel might be. “The notion of trying to superimpose over this structure [i.e., the adventure game] a dramatic conception other than a puzzle was apparently too much for the audience.” (Interviews with Contemporary American Science Fiction Writers, 1990). The poet Robert Pinsky was more obliging with puzzle-based play, as we shall see.

“Once I was deluxe; now I am debris” (Adrian Belew). 1980s graphics are often crude and few players can tolerate them, so that these 600 or so games lie in almost total neglect today and even Infocom's four late graphical titles are made playable on modern machines more out of piety than appreciation.4 Nevertheless, graphical adventures were once formidable rivals and the idea that graphics should voluntarily be given up required selling. “You'll never see Infocom's graphics on any computer screen,” boasted early advertisements, making a virtue of necessity: the games were supposedly too evocative and too cerebral, as was implied by the photograph of a brain. It would be truer to say that Infocom's graphics were instead in the highly crafted and colourful booklet built into each game's box lid, which often contained clues (partly as an anti-piracy measure). Inspired perhaps by the success of Kit Wright's book of paintings Masquerade, which encoded the location of a buried golden hare, Acornsoft offered a prize for the first correct solution of ‘Castle of Riddles’ (Peter Killworth, 1984) which – it was laid on pretty thick – was so fiendish that the prize would take quite some winning. A similar prize for ‘Eureka’ (Ian Livingstone, Domark, 1984) offered a whopping 25,000. Another strategy was to extol the sophistication of a parser, and how very far it soared above the mulish ignorance of a shoddy two-word job. Thus Magnetic Scrolls's ‘The Pawn’ was sold partly on its celebrated claim to understand “use the trowel to plant the pot plant in the plant pot”. Publicity for Melbourne House's ‘The Hobbit’, a huge success in Britain, stressed that only the presence on the programming team of an expert in linguistics (Stuart Richie) had enabled the invention of “Inglish”, as the parser's subset of English was called. In a similar bid for dignity, Infocom soon distanced itself from the boy-scout “adventure game” and the nerdile “computerized fantasy simulation”, instead billing its product first as “Interlogic” (1982) and then “interactive fiction” (1984), which remains the preferred euphemism today.

Though cave games became old hat, more cohesive fantasies and overt miscellanies and treasure hunts had continued unabated as the mechanics of ‘Advent’ lived on in the guise of different genres. John Laird's ‘Haunt’ (begun 1978, and known to the authors of ‘Zork’) may in fact be the first non-cave game, but its vampire-haunted house has ‘Advent’-like puzzles based on combining objects (throw turpentine on a poor painting to reveal a Rembrandt). ‘Haunt’ is not inspired, but the knockabout style and the unexpected arrival of James Watt from the Department of the Interior wanting to buy the house for $10,000,000 liven things up. Camp horror-movie settings, usually featuring Dracula or Frankenstein and set in large houses even more haphazardly stocked than caves (Chris Gray's ‘Mansion’ (1980) somehow works in a submarine), vied with science fiction and spy thrillers as the most popular variations. Alien worlds and derelict spacecraft, caves of steel so to speak, initially lent themselves to works of fairly high seriousness such as ‘Starcross’ (Dave Lebling, 1982), which is heavily indebted to Arthur C. Clarke's Rendezvous with Rama (again) and to Larry Niven's Known Space stories, from which it borrowed red and blue stepping discs. Level 9's ‘Snowball’ (Mike, Nick and Pete Austin, 1983), set on an interstellar colony ship, and Peter Killworth's ‘Countdown to Doom’ (1984) for Acornsoft and Topologika, set on Doomawangara (a hostile planet, not a region of the Australian outback), are equally steeped in golden-age science fiction.

More often, the future became a vehicle for comedy, usually in the form of sending up or camping up traditional science fiction, with one notable exception. Douglas Adams's radio series and novel The Hitchhiker's Guide To The Galaxy, the Three Men in a Boat of the 1970s, achieved enormous success. To computer hackers it became a devotional text, like the Monty Python sketches it owed considerable debts to,5 especially in Adams's invariable practice of having the “straight man” in any conversation argue back. For all its playful anything-goes imagination, it is not a send-up but a genuine work of science fiction in the sense of social analysis: Adams mocks something large enough to be worth mocking, i.e., real life, rather than pulp sci-fi and flying-saucer movies. With an electronic, interactive encyclopaedia as narrator and an author fascinated by textual gadgetry, Adams's comedy was a natural for adaptation to adventure-game form and his collaboration with Steve Meretzky at Infocom produced their bestselling title (1984). Imitations became commonplace.

It is often forgotten, because this is not how we think of “classic text adventures”, that many early games were earnestly or drably serious in tone. ‘Advent’ itself contains relatively little humour, despite one comedy room (Witt's End) and infrequent moments of drollery from the narrator:

A glistening pearl falls out of the clam and rolls away. Goodness, this must really be an oyster. (I never was very good at identifying bivalves.)

The distinctive comedy running, sometimes only as an underground stream, through all of Infocom's games has its source instead in ‘Zork’ which, partly because it contains many more inessential responses than ‘Advent’ (responses, that is, which a player winning the game need never see), gives a much stronger impression of personality: shaped by ‘Zork’, by inheritance of parsing code from one game to another and by a shared in-house testing team, the strongest unity of style between the Infocom games is that they seem told by essentially the same narrator. This is a theme that will recur in §48.

With a growing catalogue in the mid-1980s, Infocom's mature style was to make conscious use of genre to differentiate its products, essential since a core audience, subscribing to a newsletter, had formed and would buy many of its titles: also, of course, for the fun of it. Any player dropped into the middle of one of ‘The Lurking Horror’ (H. P. Lovecraft horror), ‘Leather Goddesses of Phobos’ (racy send-up of 1930s space opera) or ‘Ballyhoo’ (mournfully cynical circus mystery) would immediately be able to say which it was. (‘TLH’, Dave Lebling, 1987; ‘LGOP’, Steve Meretzky, 1986; ‘BH’, Jeff O'Neill, 1985.) Infocom also evolved a police procedural “detective novel” format rarely used since, which differed from any number of spy intrigues in that it involved character interaction and a developing case, rather than simply an ‘Advent’-style exploration in which an enemy base replaced the caves. Today's designers are not always so definite in keying a game to an established genre of fiction, but the first decisions remain to choose the style, the mood, the character of the protagonist and above all the fictional world of which the story itself will remain only a part.

Infocom achieved popular success in the mid-1980s and continues to be highly rated now, but to some extent for different games. The 1984 and 1985 bestsellers, ‘The Hitchhiker's Guide To The Galaxy’ and ‘Wishbringer’ (Brian Moriarty) are now found solidly mediocre, charming but insubstantial. Critical acclaim flickers instead to ‘Trinity’ (1986), as lightning flickers to an aerial, because the game opens with something of the mood of an art-house movie, because it is bookish and purposeful – a research bibliography is supplied – and because it is obtrusively trying to be what today's critics most wish to find: literature. Here is Brian Moriarty, a self-analysing and intensely driven designer, ambitious for worldly success but describing himself in the credits as a member of the Nathaniel Hawthorne Society, and a man who remains a name in the games industry today:

I amassed a pretty substantial library on the history of the atomic bomb … I went to the Trinity site [of Oppenheimer's first test] itself, visited Los Alamos and a lot of museums and I talked to a couple of people who were actually there … I wanted people, when playing the game, to feel their helplessness. Because that's what I felt when I was reading and talking to these people and seeing these places. You could just feel the weight of history on you.

‘Trinity’ is not altogether dark, nor innovative in its basic mechanics, with an extensive slice of ‘Zork’-like terrain and some mischievous animals. A black cover with mushroom cloud wrapped a typically deluxe and supportive book-shaped package which included a map of the Trinity test site, a sundial, instructions for folding an origami crane and a spoof 1950s The Illustrated History of the Atom Bomb comic for boys (by Carl Genatossio). Sales were tepid at best, albeit in part because the game played only on larger and therefore less widely owned computers. Had it sold no further even than to his mother, though, Moriarty's reputation would still rest secure upon ‘Trinity’ today.

All the same, for the origins of the deliberate artistic statement in interactive fiction one should look further back. Mike Berlyn had been instrumental in subverting the genre's initial “puzzles for treasures” definition, and the closing scenes of his ‘Infidel’ (1983, with Patricia Fogleman), re-enacting the Egyptian Book of the Dead, are arguably the earliest to be guided for consciously literary ends. Where ‘Infidel’ is clearly a plotted novel, ‘Mindwheel’ (Robert Pinsky, coding by Steve Hales and William Mataga, Synapse, 1984) is a puckish dream-poem, Dante meets Alice Through the Looking-Glass. The protagonist wakes to find that one Doctor Virgil is guiding him to meet some popular entertainers (rock singers and baseball players), some chatty, debauched insects and “your broccoli-coloured companion”, a frog who keeps up self-deprecating chatter throughout the game:

“Again!” says the frog. “Again we're in a situation of this kind. And I turned down a nice job, summer replacement for the little dog in Monopoly!”

But there are also transfigured victims out of Bruegel or Bosch, such as the children with the heads of birds who police the snowy, complacent mind of a Generalissimo. Opposition is provided by the more playful Spaw, a demon wearing “lawyerskin boots”. Puzzles include sonnet-writing and an end-game based on human chess with different puzzles on each square, the player advancing pawn-like (there are several possible paths, since captures can be made) to the eighth rank. Pinsky – noted for semi-formal verse with a social aspect – was named the US Poet Laureate in 1997, and what he now calls his “computerized novel” is not easy to dismiss. ‘Mindwheel’ was the first of four nightmarish games by Synapse Software: ‘Breakers’ (Rod Smith, Joe Vierra and William Mataga, 1986) is a tale of indentured labour with “coercive interrogation techniques” and the starship ‘Essex’ (Bill Darrah, 1985) lends a similar mood of persecution. Only ‘Brimstone: The Dream of Gawain’ (James Paul, David Bunch and Bill Darrah, 1985) – doubly unreal as a dream by an Arthurian knight – resembled ordinary adventure fodder. The line was terminated by Broderbund in 1986, which had bought the company but was now alarmed by a rape scene and some Chinese black magic. The computer games audience was sliding downward in age.

Infocom's intention to explore byways of the new medium was genuine, but not of course altruistic, and its business history throws a good deal of light on its decisions. The extent of Infocom's commercial success is often exaggerated, not in its scale (at one time a quarter of U.S. homes owning computers had bought the product) but in its duration. Typical sales per new title rose from 10,000 in 1981 to 50,000 in 1983–6, falling below 20,000 again in 1987–9. The exceptions were the ‘Zork’ trilogy, which sold 1,000,000 units over the decade – which explains if not excuses the later sequels – and ‘The Hitchhiker's Guide To The Galaxy’ at 250,000, which explains Infocom's eagerness to write ‘The Restaurant at the End of the Universe’. Sales were further buttressed by customer loyalty, carefully nurtured by large direct mail shots (at end of 1986, circulation of the newsletter was 240,000); by repackaging of 1980–2 titles; and by a no-returns policy in distribution (ended in 1987) obliging shopkeepers to treat Infocom's wares as luxury goods, kept on shelves until they sold. Remarkably, ‘Suspended’ (1983), not an obvious money-spinner, was to receive a Gold certificate for 100,000 sales in 1986: typical shelf times today are measured in months or even weeks. Infocom's customers were, according to market research, adult (75% over 25) – which is not so surprising given prices of $40 to $50 – and heavy readers, 80% of them men, though specific products were designed to appeal to women (such as ‘Plundered Hearts’ and the mysteries) and to children (Stu Galley adapted the ‘Seastalker’ parser to children's sentence structures, observed during testing). The work force had grown fast (1981, two; 1982, four; 1983, twenty; 1984, fifty; 1985, one hundred) but was increasingly preoccupied with managing itself and with Infocom's one business product, the database ‘Cornerstone’ (1986). It was intended to capitalise on Infocom's expertise in virtual machines, which allowed large programs – adventure games – to run on a variety of different designs of small computer: but this was not the strength in 1986 that it would have been in 1982, since the IBM PC had grown in capacity and cornered the business market, most of the rival manufacturers having gone bust in 1983–4. ‘Cornerstone’ sold 10,000 but only after a price reduction from $495 to $100, and by then Infocom had turned the corner into loss. In June 1986 Activision had bought Infocom, in what amounted to an agreed merger, for stock valued at around $8 million: at about five years' gross income, this was a high price, or would have been if the stock had in fact been worth that. Infocom still had fifteen titles ahead, including a few of its best, but disputes over branding, marketing and the division of profits and losses produced disquiet, as did a time-consuming lawsuit about the state of the books when the company changed hands; while Activision had its own travails. Expected sales from the Hitchhiker's sequel ‘Restaurant’ were an essential part of the business plan every year from 1985 to 1989, while Meretzky, Lebling, Jeff O'Neill and Amy Briggs were each briefly in the frame as the unlucky programmer. The project was stymied in 1985–6 by Douglas Adams's inability to get out of the bath when copy deadlines loom – “you can't fault him for personal hygiene in a crisis” (Geoffrey Perkins) – and by 1987–9 meant impossible collaborations with the British firm Magnetic Scrolls and other intermediaries, whom the Implementors were unable to establish working relationships with. Games by out-house employees got a little further: though Berlyn's came to nothing, Blank wrote ‘Journey’ from California and newcomer Bob Bates designed ‘Sherlock’ and ‘Arthur’. (Bates worked from his notionally independent company Challenge, but its finances were at that time heavily dependent on advances from Infocom.) Without conscious decision, Infocom was becoming a commissioning house rather than a workshop. The testing department was involved so late on that the new management saw it as largely an obstruction. Artistic collapse came in 1988, when four of the six remaining creative figures were fired or felt unable to go on (editor John Prince, tester Liz Cyr-Jones, Implementors Jeff O'Neill and Amy Briggs). Meretzky and Lebling remained, sometimes despondent, sometimes cheerful, doing largely terrible work. The weekly game-design lunches became at the last a charade, attended by random managers whom they barely knew. Infocom never went bust as such, but by 1989 market conditions would have obliged any management to salvage the Infocom brand-names while abandoning text for largely graphical games. The company now called Activision (following a second, happier merger) did just this with a fresh generation of game designers in the 1990s. For all that they were doing something quite different and thousands of miles away, many had a keen sense of standing in an Infocom tradition. Activision's omnibus 1990s reissues of the Infocom text games achieved unexpectedly high sales, to everybody's pleasure.

§46.3   The growth of a community c. 1985–91

One by one, the companies ceased to publish text-based games: Adventure International (1985), Synapse (1986), Infocom (1989), Level 9 (1991), Topologika (1992), Magnetic Scrolls (1992). The last stalwart, Legend Entertainment – which had inherited two of the designers of Infocom's last days, Bob Bates and Steve Meretzky – made the last mainstream release of a game with a parser in 1993 (‘Gateway II: Homeworld’, by Mike Verdu and Glen Dahlgren) and so, according to some, the dark ages began: the adventure games which flourished in the marketplace were fiction of a kind, and steerable, but no longer interactive in any conversational sense. Yet the same advance of technology which drove the irresistable rise of graphics and animation also brought interactive fiction writing into the home. Between 1980 and 1990 the personal computer went from being barely able to play a text game to being easily able to compile one.

To be able to program a PDP-10, as Crowther and Woods had, was a professional qualification, but early microcomputers came with the easily learned programming language BASIC built in, and also with manuals which emphasised that computers were more for writing your own programs than for using other people's. Minimal adventure programming is not complicated – indeed adventure game-writing has been used to teach children about computing (Creating a Database, 1985, Steve Rodgers and Marcus Milton) – and the slowness of a program written in BASIC does not much matter for a text adventure. So, parallel to the commercial market and at the cheaper end merging into it, hobbyists had had been devising their own adventures since at least May 1979, when Lance Micklus published ‘Dog Star Adventure’ in a computing magazine. This was the first of many type-it-in-yourself games which some readers, at least, adapted and reworked as they typed. Scott Adams's ‘Adventureland’ source was itself published in Byte (1980) and, within a year, reworked into the core of Brian Howarth's rival ‘Mysterious Adventures’. (No less than three Scott Adams-format adventure-making programs existed for the TI-99/4a microcomputer alone, and Adams-format games in circulation outnumber his official titles by at least three to one.) Dozens of books with titles like Adventure Games For Your Commodore 64 (Duncan Quirie, 1984) consisted of little more than unannotated and – partly because of the need to save every possible byte of memory – almost incomprehensible BASIC listings, often rushed into print by coders with more enthusiasm than skill, and seldom properly tested.

The major companies each had in-house systems for designing adventure games (see §41 for a sample of Infocom's ZIL): these never emerged into the public eye. But a popular design tool called The Quill (Graeme Yeandle, 1983), running on the Sinclair Spectrum and Commodore 64 microcomputers, allowed many hobbyists working from home to sell their wares. Yeandle's generous acknowledgement in the manual that the system “has its origins in an article written by Ken Reed and published in the August 1980 issue of Practical Computing” is further testament to the influence of 1979–80 magazines in spreading the word. At least 60 commercial releases in the period 1983–6 were Quilled. (Graphic Adventure Creator (Sean Ellis, 1986), was also popular later on. Two corresponding American commercial systems, Adventure Writer and Adventure Master were less fruitful.) Yet tiny BASIC games were inevitably toy models of the real thing, and even The Quill could not then build a port of ‘Advent’ (though extensions of it, such as the Professional Adventure Writer (1987), later did), let alone an Infocom-scale game.

Some 33 non-commercial design systems are now present at ftp.gmd.de, though some have fallen into disuse, some are flimsy at best and others are exercises in writing a Pascal-like compiler, or a LISP-like compiler, which pay too much attention to syntax and do not engage with the real issue: the world model and how designers can work with and alter it. During 1995–9, only two systems have been widely used: the Text Adventure Development System (or TADS, Mike Roberts, 1987) and Inform (Graham Nelson, 1993). A further two retain interested minorities and remain in the running: Hugo (Kent Tessman, 1994) and revamped forms of the Adventure Game Toolkit (AGT, David R. Malmberg and Mark Welch, 1985–7). Note the dates, implying both the durability of a capable system and the difficulty in getting a new one off the ground. All four have been continuously developed since their inception, to some extent bidding each other up in richness or complexity.

Design systems of the 1980s had the consistent ambition of easing the way of the neophyte programmer, who would if possible never be asked really to program at all: only to supply textual descriptions, in effect filling in a database. “Alan is not a programming language. Instead Alan takes a descriptive view,” says its manual (by Thomas Nilsson), and the Generic Adventure Games System manual (Mark Welch) concurs:

It [GAGS] cannot be used to write an adventure game with as many complex features as Infocom's. To do so … would require adventure-game writers to learn a very complex set of rules.

This is the bargain that was, with some reluctance, accepted around 1992 – a crucial year, as we shall see.

Changing conditions of computer networking have, throughout this story, had greater effect than the changing technology of the computers themselves. In the late 1960s and early 1970s, tools such as Unix and games like ‘Spacewar!’ spread through outposts of the early Internet much as, in the dark ages, classical texts flowed down monastic libraries along the Rhine and the Loire, always subtly rewritten until it seemed no definitive version remained. (We have already seen what this did to ‘Advent’.) By the mid-1980s, universities and institutions across at least the Western world were securely networked, but the same could not yet be said of home computers and small businesses.

Enthusiasts for writing interactive fiction could achieve little until there was enough mutual contact for non-commercially distributed design systems and games to spread around. Usenet, a wide-ranging system of discussion forums or “newsgroups”, was created in 1979 and roughly doubled in usage each year: but in 1985 it still only accounted for 375 postings per day, across all topics combined, and as late as 1989 it was still possible for a single person to skim through the entire traffic. In America, the hundreds of local dial-up bulletin board systems (BBSs) and their big brother, the American CompuServe service, had greater effect. Much as the creators of Usenet called it “the poor man's ARPANET”, bulletin boards were a poor man's FTP: allowing downloading of files from archives with only a modem and without the need for a university or corporate Internet connection. Even as late as 1993, downloads of ‘Curses’ (say) from CompuServe rivalled those from ftp.gmd.de. From 1982, when the concept of “shareware” was invented, quality home-made software was routinely “shared” on these boards and via discs ordered from the Public Software Library, subject to a moral obligation for users to “register” by paying the original author a small fee. TADS and some of the better early TADS games were administered as shareware by High Energy Software, a company with its own BBS. AGT was also available from the BBS community, and the Byte Information Exchange, and as discs which could be ordered from Softworks, a member company of the Association of Shareware Professionals.

Judith Pintar's story (from an interview in XYZZYnews 11) shows how fruitful this contact could be:

I started writing IF in the mid-80s, when the XT I bought happened to have GAGS [1985] on it … When I joined CompuServe in 1990, I tried to find Mark Welch, to register GAGS, and discovered that it had become AGT and was administered by the co-author, David R. Malmberg. He had run several annual game-writing contests, and I was determined to enter … ‘CosmoServe’ [1991] tied for first place. [In 1992] I had the idea of writing a game as part of a group venture. I posted this idea in CompuServe's Gamer's forum, and a small group of people responded. … We were given a private area on the forum to post our messages to one another and to share game files.

‘Shades of Gray’ (1992), by Mark Baker, Steve Bauman, Elizabeth Ellison, Mike Laskey, Hercules, Cynthia Yans and Pintar herself, duly won. The AGT contest began as a GAGS contest in 1986, to promote interest, then ran annually 1987–93 and, though no longer tied to any game design system, essentially resumed in 1995 as the rec.arts.int-fiction competition.

§46.4   Newsgroups and revival 1992–

The rest of the infrastructure of the present interactive fiction community was created by four almost simultaneous events: first, the creation of specific Usenet newsgroups (on or before 21 March 1992, 21 September 1992), moving away from sporadic and easily drowned-out talk in the early net.games and the later rec.games.programming, while asserting an artistic medium by moving to the rec.arts.* sub-hierarchy, where theatre and the novel are also discussed. Next, the founding of the interactive fiction archive at ftp.gmd.de by Volker Blasius and David M. Baggett (24 November 1992), the release of TADS 2.0 (6 December 1992), significant because it established the dominance of TADS, and the release of Inform 1 (10 May 1993), though not until 1995 was Inform seriously used.

The proximity of these dates is no coincidence: they followed the sudden, widespread and cheap release of the entire Infocom back catalogue, in two volumes, Lost Treasures of Infocom I (January 1992) and II (July 1992), which stimulated a revival both of the cult of Infocom and of interactive fiction in general. If Infocom is to be compared with Shakespeare then this was the First Folio. Anybody who had occasionally liked Infocom's games in the past suddenly had all of them, while players' expectations of quality rose. Infocom antiquarianism occupies much of the early newsgroup traffic; much of the initial stock at ftp.gmd.de consisted of fact-sheets on Infocom story files; and the collectors of these diamonds were provided with a rhinestone machine when Inform, which compiled Infocom-format story files, appeared. Design systems successful in the mid-1980s, which were not well-adapted to build Infocom-scale games, quickly died out. Instead TADS and Inform were used for a clump of large games plausibly imitating Infocom's production values, among them ‘Save Princeton’ (Jacob Solomon Weinstein, 1992), ‘Horror of Rylvania’ (Dave Leary, 1993), ‘Curses’ (Graham Nelson, 1993), ‘The Legend Lives!’ (David M. Baggett, 1993), ‘Theatre’ (Brendon Wyber, 1994), ‘Christminster’ (Gareth Rees, 1995), ‘Jigsaw’ (Graham Nelson, 1995), ‘Perdition's Flames’ (Mike Roberts, 1995) and so on.

The revival by Gerry Kevin Wilson of an annual game-designing competition, fondly remembered by AGT and CompuServe users, took place in September 1995, though the 1996 event (with 26 entries) marked the beginning of its real importance. There was no restriction to any specific design system, as was typical of a newsgroup which has consistently voted against dividing itself off into subsections such as comp.lang.tads or comp.lang.inform. However, the rule that a contest game should be solvable in two hours, albeit often more honoured in the breach, has had a decisive effect in diverting designers from Infocom-sized “novels” into short stories. This freed up the form for greater experimentation, but meant that few large games were created in the late 1990s. The annual September event – shrewdly timed, after long university vacations – has also had the unfortunate effect that games are held back and then all released at once, in the breaking of a monsoon after a parched summer. But for the quantity of fine work stimulated, and the number of newcomers attracted, the competition can only be considered a triumph. Its success has also spun off a number of alternative contests and forms of recognition, such as the annual and Oscar-like XYZZY Awards. The most ludic, madcap events are the SpeedIF contests, begun by David Cornelson in October 1998, in which just two hours are allowed to write a complete game. The rubric for SpeedIF 13: “The game will take place in a Chinese restaurant. It will feature one or more of the following animals: pigeon, elephant or badger. There will be some kind of sculpture made of mud, and some character will be obsessed with either HAL or Doraemon, the robot-cat from the future (or both)…”

A growing appreciation of the medium's potential for art has characterised turn of the century interactive fiction. One provocative example was Nick Montfort's showing of a hardback edition of ‘Winchester's Nightmare’ at Digital Arts and Culture in Atlanta, in October 1999: ten decommissioned laptops converted to run an Inform game. Formerly property of the Internal Revenue Service, several still bore the U.S. Treasury seal of an eagle holding a key. Even shown in images on a web page, it was startling as a work of conceptual art, and challenged the unconscious assumption that an interactive fiction need be intrinsically unlike a material book. Chris Klimas's affecting short story ‘Mercy’, and Andrew Plotkin's shifting vignette ‘The Space Under The Window’, were influential in the late 1990s style for non-game games. Marnie Parker organises art shows which challenge traditional aspirations for IF by encouraging artistic expression. Making an exception to this chapter's general rule of stopping history at the close of 1999, it seems appropriate to finish with two fine debut pieces which took awards in the 2000 Art Show: ‘Galatea’ (Emily Short, 2000), a conversation with an animated sculpture which breaks new ground in interactive dialogue; and ‘The Cove’ (Kathleen Fischer, 2000), an evocative seascape which is also a gathering of memories. Interactive fiction will always appreciate what in theatre used to be called “the well-made play”, the polished entertainment on traditional lines, but without its radicals it will die. Though the grail of puzzle-free yet interactive literature seems as elusive as ever, it is too soon to stop looking.

And what of the history of the theory of interactive fiction? For most of the last twenty years, the best published sources are chapters which, like this one, hide at the back of books on game programming: thus, Chapter 8 of Peter Killworth's How to Write Adventure Games for the BBC Microcomputer Model B and Acorn Electron (1984), Chapter 7 of the Alan manual, Appendix B to the TADS manual. (Even people who don't intend ever to use TADS should read the delightfully written TADS manual.) Most are couched in the form of avuncular advice (“No matter how small an Adventure you write, it will take far, far more time and effort than you thought it would” – Killworth) but there are often clear signs of groping towards a systematic critical model of what the essential ingredients of a game might be. Basic Adventure and Strategy Game Design for the TRS-80 (Jim Menick, 1984), a turgid work, talks about “layering” complications into “phases”. Others, such as A. J. Bradbury's book, quoted above, or the somewhat gauche Player's Bill of Rights (Usenet posting, 1993), boil opinions down to (usually ten) golden rules. Bradbury's are well-supported by argument and engage with the underlying fiction and not simply the surface puzzles: for instance, “resist the urge to create a superhero” of the game's protagonist.

For interactive fiction history, see the Infocom Fact Sheet by Paul David Doherty, the Level 9 Fact Sheet by Miron Schmidt and Manuel Schulz and the Magnetic Scrolls Fact Sheet by Stefan Meier and Gunther Schmidl. Hans Persson's Adventureland catalogue and the master index of ftp.gmd.de are also indispensable. On connections with literary precursors, see Gareth Rees's 1994 article Tree fiction. The number of games said to have origins before 1979 is somewhat akin to the number of American families claiming descent from the Mayflower pilgrims and one must approach claims to priority with caution. But it is clear that too little is known about the games libraries in circulation in the mid-1970s. Peter Langston's ‘Wander’ (1974), a text-based world modelling program included in his PSL games distribution for Unix and incorporating rooms, states and portable objects, was at least a proto-adventure: perhaps many others existed, but failed to find a Don Woods to complete the task? So much appears lost that even Crowther's original source code, the most important document we might want to see, appears not to have existed anywhere since 1977. (Crowther confirms that he now has no printouts or notes of it.) In the discussion of authorship above I have therefore relied on anecdotal accounts in print quoting Crowther and Woods, and on recent and valuable research by Dennis G. Jerz.   The Longest Cave, by Roger Brucker and Richard Watson, includes a history of the Bedquilt Cave.   It is difficult to estimate the extent of the literature with any reliability. The ftp.gmd.de archive contains over 1,700 games which have been finished and offered to some kind of audience, and this necessarily excludes some material still under copyright or simply lost. Hans Persson's catalogue (surely incomplete) lists around 800 commercial publications in the 1980s, of which 330 are purely textual. The well-established canon of “important” or “classic” games still talked about today consists of about 100 titles. At most fifty titles of the 1990s remain in regular play through being rediscovered.   The Quill and its variants continued even into the 1990s, at least for the prolific Dorothy Millard, whose games are off-beat variations of “you are stranded”: the most off-beat of all being ‘Yellow Peril’ (1994), in which the entire world has become yellow.   1993 was the year of the explosion of the World Wide Web, during which it grew by a factor of 3,000. But in 1992, personal web pages barely existed, and it seemed not only natural but unavoidable to house games at a centralised library, ftp.gmd.de. The continued importance of this IF-archive is a major part of the solidarity of today's community.

1  Queneau's novel Zazie dans le métro is freely adapted, and translated into Italian, by Luca Melchionna's Inform game ‘Zazie – Una lettura interattiva’ (1999).

2  ‘Journey to the Center of the Earth’ (1978), by the remarkable Greg Hassett – then aged twelve – had better be disqualified, unless Jules Verne's original really has a coke machine, a troll's palace and a car repair shop.

3  Thus A. J. Bradbury, in Adventure Games for the Commodore 64 (1984, with variant editions for other microcomputers), a wide-ranging and thoughtful book: for instance Bradbury discusses defects in ‘Philosopher's Quest’ (arbitrary) and ‘The Hobbit’ (too little thought to different possible orders of play – perhaps this is why many winning lines score more than 100%), advocates mapping on a linked-octagon grid and so on. The magazine alluded to is Micro Adventurer, edited by Graham Cunningham and published by the small press Sunshine Books. It ran for thirteen issues from November 1983, moving gradually away from snippets of program listings as reviews and general articles took over.

4  While Spectrum and early PC graphics are often unbearable, the Atari ST and Amiga offered a better colour palette and resolution: in these versions, the Magnetic Scrolls games were well illustrated. Of lower-resolution games, ‘Asylum’ (William Denman), ‘Sherwood Forest’ and ‘Masquerade’ (Dale Johnson) are recommendable.

5  ‘Advent’ quotes from Python's Parrot Sketch: try feeding the bird. As for Adams, ‘Lord’ has a babel fish and ‘Acheton’ a ningy.