This is a great article -- but (some) commenters are missing that "type-in text games" means two different things in context: text games are games whose interface is primarily text-based, but type-in games here means "games distributed as source code printed in magazines or books". This is something that basically disappeared from computing history by the mid-1980s, as it stopped being even remotely practical as a distribution method.
This article, by the way, is sort of bonus material from Aaron Reed's 50 Years of Text Games, which is phenomenally interesting if you're the sort interested in that corner of computing history. While many "text games" are parser-based like Infocom's text adventures, Reed also gets into Twine games, Fallen London, Dwarf Fortress, King of Dragon Pass, ARGs and more.
I have collection of old magazines and just looking at pages of possibly even compressed code makes me somewhat miffed... They even added check sums on rows later to make it simpler process.
Whole thing was going to die with home computers, programs just grew too much for it to actually work.
I remember once in the early 80s typing in a very lengthy program listing from Byte or Nybble magazines on an Apple ][e.
It wasn't BASIC or even Assembler, but machine code!
It didn't run at first so I printed it all out (on 15" wide green bar paper) and on several bus rides to and from school would compare the magazine code to my own print-out until I found the few bytes I had mistyped. Checksums would have been nice!
IIRC it was some kind of music program that let you type in melodies and play them back on the PC speaker. I took a melody line and a bass line from a song I was writing at the time and typed them in on two neighboring PCs in the university lab. When all was ready, I pressed the two Enter keys at the same time, and after weeks of work had several seconds of glorious full-volume square-wave harmony! Formative moments!
> This is something that basically disappeared from computing history by the mid-1980s, as it stopped being even remotely practical as a distribution method.
It was very much alive as the distribution method for getting TI-BASIC games onto calculators in the mid-to-late 90s.
If you didn't know anyone with access to the parallel-port graphlink cable, your only option was transcribing it manually. One someone transcribed a copy, distribution with the included male-male cable was possible.
I had the book "Cosmic Games for the Vic 20". I tried to type in some games, but wasn't all that impressed. I think it was "Treasure Hunt" and the bird broke the newline, leaving the screen looking off.
I learned English as a second language from playing Leisure Suit Larry in the Land of the Lounge Lizards when I was a kid.
Didn't know how to pronounce "Suit" until I was about 18.
To this day I am disappointed by the "Point & Click" remakes. It was probably the first hint that society as a whole develops in ways that does not always align with my own interests.
Building a story out of text, as opposed to visual imagery, allows for a different kind of phantasy. I like both, and I sure miss the amazing feel of interactive text and graphics based adventures that Sierra produced in its early days.
> To this day I am disappointed by the "Point & Click" remakes. It was probably the first hint that society as a whole develops in ways that does not always align with my own interests.
I'm torn by this. I played a lot in the 80s and 90s, and mostly took a hiatus till very recently. I too noticed the games are much easier - even way easier than the 90's point and click adventure games.
However, I also realize that I'm now much busier, and if they were as challenging as the older games, I simply wouldn't be able to play them. I do appreciate that easier games now exist for people like me!
The other thing I noticed: The storytelling in modern games is significantly better!
People still make interactive fiction games that are solely text (think Infocom style), and from what I've heard they are superior to the old commercial ones. So there's always that.
Same here. English was my third language which I absorbed mostly from Sierra games at age 9-12. I didn’t speak it with anyone until years later.
The Larry games certainly offered some interesting colorful vocabulary and cultural references which all flew over my head. The word “prophylactic” still instantly makes me visualize the shopping scene from the first game.
The funny thing about learning a language this way is that English is forever stuck as a written language in my mind. I can speak it just fine [1], but it feels somehow a bit wrong. I’m guessing medieval clergy might have had a similar relationship with Latin.
- -
[1] Except for the pronunciation of the word “pear”. I still can’t remember if it rhymes with fear or bear. To avoid embarrassment, I stay away from food products containing this fruit.
broke that rule, I think the creator wanted to fit every idea he could into the first game. Ultima was vastly influential on western CRPG's but I think it was even more influential on JRPG's like Phantasy Star (which has that genre mixing), Dragon Quest, etc.
which David Ahl wrote to help promote DEC's minicomputers which I saw a lot of as a kid because I grew up in New Hampshire right next to DEC's headquarters in Massachusetts and DEC donated numerous machines to schools. Before my high school got a VAX (a 32-bit machine a lot like the 386) it had a PDP-8 which had a printing terminal and two video terminals and I think 24K 12-bit words. It could support three sessions running BASIC or boot up in a one session mode and run
Some schools had much nicer PDP-11 machines which had 8-bit words and could give each user a 64-bit address space so the BASIC experience was a lot like a top-of-the-line home computer. Dave Ahl saw micros were the future so he started Creative Computing magazine and came out with this sequel in 1980
I had a TRS-80 Color Computer back then which was pretty unusual and my take was that most games from those books were pretty portable and you could usually get them working on any micro of the day with simple modifications if any.
Now those text adventures were a special case, there is the issue of (1) not wanting to reveal the text, but also the fact that (2) is is cumbersome to express that kind of behavior in basic, and (3) the programs get very big. Text adventures of the 1980s were usually written in a specialized object-based interpreter. The most famous was infocom's z-machine
developed his own interpreter that was simpler but more specialized than the Z-machine. Scott developed a BASIC version of his interpreter which you could type in, and also developed assembler versions of the interpreter for many machines like the black&white TRS-80, Apple ][, etc.
Both the ZIL machine and the Scott Adams Interpreter had compilers which would turn human-readable code into a binary file that the interpreter would play.
For what it is worth, there is an electronics teaching lab at Phillips Hall at Cornell University named after David Ahl that I often walk past on the way home from work.
Took me a while to realize that this was not only a game which you interacted with by typing commands, but also a game which you actually had to type in. I was aware of the type-in programs published in 80s computer magazines, I actually typed a few of them into my Atari 800 back in the day, but didn't know there were also books containing type-in programs.
I started porting Super Star Trek from "BASIC computer games" to the web browser, but took several artistic liberties and now it's a completely different thing [1]
> The Antagonists source listing in fact consists of multiple programs: a decoder to error-check the compressed listing and convert it into assembly code, which included a virtual machine for interactive fiction; a block of data for that VM to read which defined the text, map, and behavior of the game itself; and a runner program to start up the VM, load the data block, and execute it. The instructions for getting this whole mechanism up and running were rather elaborate, with a series of procedures that had to be followed exactingly for the game to eventually run.
It's somewhat reassuring to have this example that shaky towers of abstractions are not exactly newcomers to software development.
Compute!'s Gazette had this down to a science. When I first started, they were offering simple BASIC type-ins, and I'm pretty sure that I did some of those without benefit of persistent storage.
But then they began trumpeting a revolution in type-ins: just enter this simple little BASIC editor, then type in a whole mind-numbing mess of meaningless numbers. The editor checksummed every line, et voila! A working and extremely performant game/app entirely in machine code!
The checksum was amazing! When my siblings and I made a mistake, we would go through the entire code and check each entry. Took too long and we gave up!
The very first thing I did at a 6 year old w/ a brand new Vic 20 for Christmas, '82, was type in a game called Parrot from a book that was bundled with it.
It literally was a "teach you to type" game. The best I can recall, letters fell from the sky and a crude parrot would squawk every time you hit the correct letter.
In a way, computer games are an evolution of interactive storytelling, so I don't see the "fall" of text-based games as a negative, but more as the result of the progression of the electronic medium as a storytelling device.
Before computer games, I remember being entranced by Choose Your Own Adventure books. Text adventure games are an evolution of this concept, and as the medium got more advanced, users/readers/players demanded more interactivity.
MUDs are still popular with those who enjoy them, just as books are still read when interactivity is not the focus. But overall, interactive storytelling has evolved in many ways from what it was capable of 40 years ago, and most fans moved on with it.
In the future, interactivity will only increase, as we move towards interacting with "real" AI characters, where the world and the story itself can be created just-in-time, versus the ahead-of-time process we use now. So the stories and experiences of each player will be unique to them, rather than preconceived by writers, designers and programmers. This is exciting, and somewhat scary, depending on how you look at it. :)
Like books, text-based games leave much to the player's imagination. There's no (or very limited) graphics, and the player's mind fills in the gaps.
Which stimulates player's creativity, and makes the game a unique experience to each player.
But there's a fine line between exploiting this successfully, and relying on it too much - at which point players get bored.
Modern games rely much less on this. At least not for how the environment looks. Instead machine works hard to present a semi-realistic environment. Enough to "suspend disbelief" anyway.
This lowers the bar for casual players to get into a game. But at the same time, makes for a (somewhat) more passive experience. You explore a pre-designed environment vs. create a good part of it yourself.
Exception might be 'builder' style games like survival games, some RTS / RPGs & of course Minecraft.
> This lowers the bar for casual players to get into a game. But at the same time, makes for a (somewhat) more passive experience. You explore a pre-designed environment vs. create a good part of it yourself.
I think this is less about aesthetics and more about directed versus undirected gameplay. Mainstream/AAA modern games more or less tell you exactly what to do and exactly where to do it, which makes for a significantly more passive experience as you don't need to think at all to progress.
An example is if you played chess with an AI telling you what moves to make. The player playing without thinking is due to a lack of player agency, not the pretty graphics of chess.
> Mainstream/AAA modern games more or less tell you exactly what to do and exactly where to do it
Just to clarify, are you stating this as a truism that I may be oblivious to, or as in your own personal experience?
Otherwise, this is proof that two different people can have such vastly different experiences. I find modern AAA games to be rich of actions and choices. There are more "open world" games now than ever. And while your statement is true of the main storyline, I think there's so much diversity in games now that the main story line is often just a "thing you get around to" after you get bored of whatever side quests or exploration you are performing.
Current Tomb Raider just requiere you to press E and most actions are quick time events.
Compared to Tomb Raider 1-5, and even Anniversary/Legend, the current AAA games are very bad.
Minecraft took me back to when I played MUSH games. You are presented with a world to inhabit, and there were no rails. You got out what you put into them game, so if you contribute little to the experience, it felt dull and lifeless. This turned lots of folks I know off from them game, where I was enamored by the possibiities, seemingly only limited by my own imagination. MUSHs were a lot like that, where MUDs were a bit more linear.
I wondervat the idea that these types of games engage a different part of the brain, similar to how some productivity coaches might suggest that writing ones goals, as opposed to talking about them, helps your brain process them in s more robust way. Citations probably needed on that, I'm no psychologist, but as someone who grew up in the MU* community and still adores text-based games to this day, I find I have stronger memories tied to those games than I do with the graphical games I've played.
While agree that some games are representing an evolution of interactive storytelling, I think a much more important aspect of computer games is being an evolution of games in general (e.g. games of skill, board games etc.).
This is in my opinion why type-in text games do not work well since they clash with what you expect from a "traditional game" with an obvious set of options for action. This is also why I think that interactive storytelling will always be only a branch of video games, never the pinnacle.
I got a subscription to Family and Home Computing about a year or two before we could afford a computer. Usually I’d throw the old issues away but one had the game of Asteroids in the back in BASIC. I held onto that one. A game I didn’t have to beg my parents to buy? Yes please.
When I finally got a computer I typed it in, which took impatient little me days to complete. And when it was done it did not work. I spot checked my transcription and couldn’t find much. I had written hundreds of lines of BASIC total in a class and this was three columns of print over something like, four or five pages. No way was I up to debugging that. One of the big disappointments of my childhood.
I still keep and maintain some text-based games. I even wrote a new one last year to further familiarize myself with golang. There's something special in them. You can imagine a full world, 3d, or otherwise. Yet you are limited to ASCII. You can get away with a lot in ascii. From Diku, to Merc, to Smaug... I try to keep the codebases building using the latest gcc.
I wonder if a modern equivalent of a game like this would sell at all. Modern supposedly story-driven games often have extremely short flavor/clue text so as not to interrupt the flow. Reading the clue docs outside of the game would allow you to enjoy the experience even through moments when you aren't playing the game itself, and it wouldn't break the pacing. I suppose the game could give you 30-40 page docs or book chapters as PDFs throughout the story as you unlock them by finding books in the game.
Wizardry 8 (2001) probably had the last remnant of a type-in interface for the communication with NPCs - you could type your lines directly and they reacted to keywords. You could ask for things you've heard somewhere and learn something new.
The idea isn't dead though - since GPT-2 appeared, people have been trying to use chatbots for role-playing games again. With GPT-4 it's almost possible as it can mostly keep track of the inventory and more or less stick to the rules.
What a fascinating trip back in time. I had almost forgotten all the time I spent typing up code listings from the back of Usborne computer books as a kid in the late 80s.
It was really a great way to learn programming, although I remember the frustrations trying to work out which variation of BASIC (Commodore, Spectrum, BBC Micro or another one) would work on my ’286 with GWBASIC.
It is interesting that reading a bunch of good books is considered a crucial part of learning to write well, but in programming classes we tend to just explain the syntax and then say “so here are some projects…”
Setting aside the type-in aspect, the idea of having a separate book (or other media) closely integrated with a game is interesting. I'm not sure how much sense it makes nowadays, since you can easily have a PDF/movie/etc on the computer as well, and once you're doing that you might as well integrate it fully into the game. Shenzhen I/O did something like this; it came with a list of datasheets for the electronic components you used in the game, and the devs recommended printing them out and putting them in a binder for added immersion. Some of the other Zachtronics games might have done something similar, I'm not sure.
The PS2/DS game Flower Sun and Rain did this. There was an available book that was a guide for an in-game resort whose contents were used to solve puzzles. There was also a digital version within the game, but the physical book looks fun.
This is a little beyond the development of text-based games, but I just want to give a shout out to Ray Dunakin's wonderful point-and-click indie creations that can be found here: http://www.raydunakin.com/Site/Games.html
I spent countless hours as a kid on "Twisted!", "A Mess O' Trouble", and "Another Fine Mess".
Some of these titles are being re-released on OSX which is exciting for me (and thank you HN crowd for causing me to investigate what happened to Dunakin's work!).
It's peculiar how you can write an article subtitled "The rise and fall of type-in text games" without even mentioning the words Sierra, Zork or Infocom.
Not sure if you read TFA, but it was about games whose code you “type in” from a book or magazine. I did it a lot back in the 80s when I first started programming.
The Sierra games took typed commands, but you got the games on diskettes. That was a later era than the games mentioned in the article.
I thought this was going to be about the inability to type in games. Hearthstone doesn't allow you to communicate with your opponents. Or trade with them for that matter
Lair of the Minotaur (proto zork) was the gateway computer drug for me. i still have an old spiral bound graph notebook from 87 which i used first to map out the dungeon and then later create my own. what a marvelous time to be a child in america
I have fond memories of King’s Quest and Space Quest on 5.25" floppies!
In the beginning of Space Quest I think you had to use a piece of glass from the broken canopy of the ship to redirect a laser that was blocking the path.
I grew up on Infocom games on apple iie. However now the focus is largely on graphics to the extreme, it could be that a lot of people simply cannot read and write well enough to get into a text based game, they had a comeback during the dialup internet days with Muds. If one went and then had such a game translated into enough languages perhaps it could find a world wide niche.
This is reaching back a bit, but Old Man Murray had a great article about adventure games that seems relevant.[1] OMM was talking about graphical adventure games, but I think a lot of it applies to pure-text games, too.
The game described in the article is something I’ve done in ChatGPT 4 months ago, its just a prompting problem, it can persist characters across rooms/scenes and keep an overarching story or goal
Any amateur IF writer with Inform6 (don't even start with Inform7, where you are basically hand-guided to do very complex tasks) can do much more than you did in weeks but without coherence issues. Ever.
Go play some decent and complex IF from Anchorhead to Spiritwrak, Curses/Jigsaw!, and you'll understand what a proper adventure and its object interaction actually are. I7 and its libraries allows to create stuff like Counterfeit Monkey, where you LLM would suffer a lot trying to implement that semantic monster. And yet, it will run in a Pentium II like nothing with a Glulxe interpreter.
Most text adventures have been curated over 30 years. From a 486, Pentium II or a tiny Rpi A/B, most top IF games from the community will surpass by large the pseudo duct-taped environment (if there is any) an LLM would create.
Zork came from Dungeon, and it's nearly 40 years. Now, the 90's. By just 1993 with Curses!, any DOS/Linux guy with a 486 could play (and still do) pretty complex games with far more interactivity than a current free roam game except for Minetest/Minecraft. And, even with MC, there are some text adventures with pure alchemy on it, allowing to craft new objects inside the game. And, I repeat, playable under a 486/Pentium without any effort.
If someone created a text adventure with the requeriments of a Pentium II/III, it could basically host a whole MUD but for a single player and tons of actions and methods to solve puzzles.
If we ramp up the requeriments for a text adventure being a Pentium 4/SSE2, you basically get Cataclysm DDA but with a text parser and 200X of stuff to do ingame, as it would not be restricted to have to render the whole world each turn. You would get procedurally generated puzzles, NPC's with personality traits, emotions, and even a relatively complex grammar on subjects. Like the IF Galatea, but in a free roam world, with NPC's interacting with each other shaping the world dinamically.
There are different aspects of how an LLM can be used; allowing the user free-input of English is clearly one way that those sorts of interfaces can be significantly improved.
Current parsers are much better than the ones from the 80's. By the 90's they were improved a lot and they will still run in a 486. For instance, most *.z8 adventures from IF Archive (the decent ones) will parse stuff like "Do X, and then do Y and Z and J it.
This article, by the way, is sort of bonus material from Aaron Reed's 50 Years of Text Games, which is phenomenally interesting if you're the sort interested in that corner of computing history. While many "text games" are parser-based like Infocom's text adventures, Reed also gets into Twine games, Fallen London, Dwarf Fortress, King of Dragon Pass, ARGs and more.