A brief history of amateur IF (interactive fiction). Heavily biased to the post-1991 period, since that’s really when I first got involved.

The scope of IF was quite limited in the early days. Adventure was born on a large time-sharing system, and only people in universities, research institutions and some companies had access to such a beast. I’m not going to comment much here about Adventure, primarily because so much has been ably written elsewhere. (most notably Graham Nelson’s excellent Craft of Adventure, now part of the Inform Designer’s Manual)

The rise of the home computer market in the late 1970s and early 1980s led rapidly to IF’s golden age thanks to remarkable output of the Infocom Imps (implementors), but actual IF development was limited primarily to the professional, commercial community through much of the decade. In the USA that meant Scott Adams and Infocom; in the UK, Level 9 and Magnetic Scrolls. This was mainly because the tools to develop IF just weren’t there. Scott Adams’ stuff was technologically primitive (great for the day, but he was working on extremely tiny computers), Infocom did their authoring work on minicomputers, and I don’t know what Level 9 and Magnetic Scrolls used. A few other commercial firms tried their hand at text adventures, but none really took off. Most companies (eg: Sierra, LucasArts) concentrated instead on graphical games.

Although the absence of good tools ruled out the possibility of amateurs with good writing skills but poor programming skills coming up with games, lots of other hacking-oriented hobbyists put together their own games from scratch. But without a decent IF application framework (there were simple systems like Eamon for the Apple // and the QUILL for the Spectrum, but nothing really that sophisticated) everyone had to roll their own parser, save-game routines, etc. And so many of those amateur games were frankly quite awful, because it’s really quite challenging to assemble such code. Also, the typical machines accessible to a home user of the day really weren’t very powerful - they were 8-bit systems like Apple //s, Commodore 64s, Ataris, BBC Micros and Spectrums. And so, to be blunt, I’m not going to focus much on this period. Partly because I don’t remember or know that much about it and partly because games of the period tended to be rather primitive and simple.

It wasn’t really until 16-bit machines - IBM’s PC and its clones - and AGT - the Adventure Game Toolkit by David Malmberg and Mark Welch - came onto the scene that amateurs really had the tools to write a half-decent game. Although primarily an MS-DOS system, there were ports to the Mac, the Amiga and the Atari ST. (sadly, the Mac port was quite flaky and only worked on the very first generation of Macintoshes.) AGT did very well in the thriving BBS community of the day, particularly the early online services like CompuServe, GEnie and Delphi. Home-based BBSs also got into the action, as you could play an online version of an AGT game on a lot of PC-based BBSs by using a "portal." Tons of games were produced during this period, and a wide selection of AGT games are still floating around the net today. There were annual AGT contests sponsored by AGT’s creators, as I recall. (the program was initially published by Softworks as shareware, though it later became freeware) And a lot of IF-related discussion went on in the gaming fora on those proprietary systems, particularly Compuserve’s. There were probably also some Fidonet echoes (an early Usenet-like system of interlinked private BBSs) on the topic. I don’t know if any records have been kept of any of those discussions - I’m sure many of them were very interesting.

AGT was popular but it, like several other development systems of the day (including AGT’s predecessor GAGS, and AdvSys by Dave Betz), really could only take you so far. You couldn’t write anything as technically sophisticated as a later-period Infocom game with it. The first really useful tool for writing amateur IF was, in my opinion, TADS - the Text Adventure Development System, by Mike Roberts. It came onto the scene in 1988, only a year after AGT, and again was distributed primarily through BBSs. In fact Mike Roberts and his business partner Steve McAdam eventually set up a small dialup BBS in Palo Alto, California, where High Energy, the company they formed to sell TADS as shareware, was based. However, TADS really didn’t take off until Dave Leary and Dave Baggett started writing the Unnkulian series of games. I think these were really the first games to show off TADS’ potential as a truly useful game-writing tool. (I don’t mean to put down Mike’s own Deep Space Drifter here, but it never garnered the kind of popularity that the two Daves’ wacky fantasy games did, and certainly wasn’t as widely distributed.) Indeed, these early TADS 1 games showed that it was possible for even a hobbyist to sit down and write a near Infocom-quality game - you didn’t need a minicomputer and vast budget to do it. You just needed a copy of TADS, with its excellent parser, totally object-oriented internal structure and extensive documentation. And just as Infocom’s games had been distributed on a wide variety of OS architectures, TADS was ported to the most common OSs of the day - MS-DOS, Macintosh, Atari, UNIX.

The release of the Unnkulian games coincided with the rise of Usenet and FTP archive sites. And it’s the Internet that has proved to be the second key element in the resurgence of interest in textual IF, after easy access to tools. Popular Macintosh writer Adam Engst sponsored the newgrouping of rec.arts.int-fiction (raif) in around 1987, although it rapidly became a forum for discussing conventional text adventures rather than more experimental forms of interactive fiction, as he’d hoped. A few years later a sister newsgroup, rec.games.int-fiction (rgif), which was formed in order to separate discussions of game playing (and requests for hints) from discussions on IF authoring.

The two groups soon became the main meeting place for a loose band of IF aficionados, and I believe that this is a really important point. Up until then IF was just too small and fragmented a market to go anywhere. It was considered commercially dead, particularly following the failure of Infocom. The field was an obscure and anachronistic specialty interest, and it took the entire Internet to assemble a critical mass of people interested in sustaining the genre. Also, and perhaps more importantly, at that time the Internet was still largely the domain of the academe. That meant that you had a lot of students wired to the net. Students possessing a crucial mix of five elements: a lot of free time, access to decent computers, an interest in geeky hackery, often an interest in fantasy or SF literature and a fond memory of playing IF as kids or teenagers.

TADS quickly became the development system of choice for the nascent IF community on raif. Conversations frequently focused around implementation questions in TADS, though of course were often related back to more theoretical questions of game design. Documents such as David Graves’ essays on IF and Infocom implementor P. David Lebling’s Byte article on the Z-machine architecture made a great deal of impact. Unfortunately, not that many games were actually released during this period, but enough people paid High Energy the $40 US shareware fee to make the enterprise worth Mike and Steve’s while, and TADS updates came regularly. Version 2, released in 1992, was another important milestone as it contained a virtual memory system that let you write very big games - larger than Infocom’s, in fact. It also boasted a comprehensive printed manual.

Another crucial milestone was reached in late 1992, when Volker Blasius opened up a section of the GMD (a German mathematical research institute in Sankt Augustin, Germany) FTP site as a repository for everything IF. This was critical because until the IF Archive came into existence there was no central point at which things could live. Usenet was still very transient back then, as archiving services weren’t widely available. You had to assume that once your post had expired on everyone’s server, it was gone forever unless you had saved a copy locally. But the IF Archive gave us a central focus for all of this stuff to live more or less permanently. A focal point from which the Internet IF community could really grow.

Then, in May 1993, things really changed. A modest and unassuming poster named Graham Nelson (who posted occasionally to raif from Oxford University under the mysterious name "Mathematical Institute") politely announced that he had created a new compiler, named Inform, that could actually create Z-machine games, the same game format used by Infocom. Until this point interest in the Z-machine was largely archaeological. A number of online hobbyists, including the group known as the InfoTaskforce (ITF), Mark Howell, Marnix Klooster, Paul David Doherty, Stefan Jokisch and Mike Threepoint, had delved into the workings of the Z-machine, documenting its inner secrets. There were also at least two programs that could decode and interpret a Z-machine game so you could, for example, play the old Infocom games from your childhood on your university’s UNIX multiuser system. But nobody had actually seemed to think of the now-obvious idea of using the same virtual machine to create new games. Or, if they had, nobody did anything about it until Graham.

The impact of Inform was slow in coming, but after an initial period of skepticism (most people actually coding stuff were TADS users at this point) Inform developed a very large following. Along with Inform Graham released Curses, the real catalyst for Inform’s success. Complex, thematically diverse and idiosyncratic, Curses firmly demonstrated that it was possible to write excellent games in Inform.

Within a couple years, Inform had more or less superseded TADS as the main language of IF discussion online. This can be attributed to a number of factors. First, and most importantly, Graham didn’t charge for Inform, whereas TADS was distributed under the shareware model. Second, the source to Inform was freely available. Graham retained copyright and did not allow people to distribute modified versions, but the guts were still there for you to look at and port to your platform of choice. Third, the Inform parser was fully accessible to the author, whereas TADS’ parser was built into the interpreter and only indirectly modifiable, by various program hooks. Fourth, Inform had that indefinable cachet of relying on the Z-machine. And a lot of people were really quite tickled to know that they were writing games that used the same internal format as that used by the early masters at Infocom. And finally, Inform’s growth coincided with a period that saw Mike Roberts taking a hiatus from updating TADS owing to other life priorities at the time. Inform’s success is also noteworthy in that, prior to the Internet, software developed in the UK tended to have very little influence outside Europe. For example, the thriving amateur UK IF world of the 1980s, based around the Spectrum authoring system the QUILL, was essentially unknown in North America.

Other developments helped continue the momentum in the IF community. In no particular order these included Kevin Wilson’s annual IF competitions and SPAG game reviews; Activision’s re-releases of the Infocom classics (which included IF competition winners on one of the discs); Eileen Mullin’s XyzzyNews online zine; Stephen Granade’s Interactive Fiction site on the Mining Co., a commercial website; Steven van Egmond’s in-depth raif archives (these predate DejaNews); Thomas Nilsson and Göran Forslund’s ALAN (Adventure Language, one of the first useful development systems to support a non-English language; Swedish in this case), Kent Tessman’s Hugo, one of the first post-TADS systems to support graphics and sound; Mike Roberts’ updated TADS, multimedia TADS, which uses HTML for markup and supports graphics and sound; Andrew Zarf Plotkin for several interpreter ports and the new Glk IF system; and Liza Daly’s ifMUD, a social gathering place for IF fans.

Um. What else? Probably other stuff I’ve forgotten here. Also, I suppose I should update it now that T3 and Glulxe and so on are out. Also talk about the release of Once and Future (Avalon), the launch and failure of Cascade Mountain Publishing, Adam Cadre’s Photopia, Emily Short’s conversation-based games and Graham Nelson publishing the DM4.

- Neil K. (tela@REMOVE-ME-tela.bc.ca)