The arbitrary landscape we stake out as Interactive Fiction looks a bit like this in 2026:

Parser games

Games with a world model, where the player interacts via verbs.

Choice games

Games primarily in text where your main interactions are making choices.

Quality-based narrative

Games in which you amass a large variety of resources (not just inventory items) that give you access to small choice-based passages called storylets.

This is neither complete nor a proper taxonomy. Broader views of interactive fiction might include: visual novels, gamebooks, tabletop role-playing games, participatory theatre, or tarot cards. Nevertheless, the above three types are quite common and a good way to start thinking about how you might approach the problem of having fiction that interacts with the player.

One way to approach interactivity is to think of it as interacting with a game world. The game simulates the world and the resulting interactions and environment do the storytelling for you. This style of interaction is great for puzzles and exploration.

The parser is the middleman between the player and the world model. The player types commands that are interpreted as actions, the action is simulated which results in textual output. In slow-motion:

1. The player types: PICK UP SWORD

2. The parser breaks the command into pieces and tries to understand the pieces via grammar: it assumes a verb of PICK UP and a direct object SWORD.

3. The game world takes that action and determines if it succeeds (or even makes sense). It simulates a successful action, such as moving the sword to the player's inventory.

4. This action is turned into some form of output: You pick up the sword.

This interactivity loop is both a blessing and a curse. There is an incredible freedom of feeling that a player could type anything in and a good game will resolve it satisfactorily. But there is an incredible onus on the developer (and player!) to make that freedom work. Typically these games will have a set list of actions that they allow so you don't have to deal with a player typing PONTIFICATE VIS-A-VIS MAN'S INHUMANITY TO MAN, unless the developer wants that in the game.

Point-and-click adventure games can be seen in this light with their set of verbs down the bottom of the screen and the ability to apply them to anything in the scene.

But even with a small set of prescribed verbs, all of the objects in the world will be fair game. The feeling of freedom, simulation and coherency falls apart after the player types LICK <object> for the tenth time and get only a boilerplate vague response.

Parser games are therefore tricky for players and developers alike, but were some of the first examples in the genre.

Modern parser games are often written in Inform 7, TADS 3, Quest or ADRIFT. That said, most of these platforms were released in the early 2000s, and not seen a huge development for several years.

Another way to approach interactivity is to think of it as a series of choices. In choice games, picking up a sword is a choice in a story. We would want to record the choice, but you don't need to model objects, inventories, containment and all the other parts of the game world simulation.

A classic version of this are Choose Your Own Adventure books. These were interactive fiction in a static medium of a physical book. You would read passages and at the end of each passage might be a choice you can make. Each choice would indicate a page to turn to if you make that choice.

Modern choice games are often written in Twine, ChoiceScript or ink. The programmatic approach allows much greater flexibility in how choices are presented and honoured. For example, a game like Detritus provides a simulation of its world. You have an inventory and explore a dynamic world. Your interactions are choosing items from a menu, but sometimes that is picking up objects from the room, sometimes it's weaving a path through narrative.

Choice games tend to be easier for people to begin developing in. Twine has easy-to-use tools that lay out your game like a flow chart. ChoiceScript allows you to write your game as a text file with intuitive formatting to provide choices.

Our last type of interactive fiction could be seen as a choice game, but it has enough design changes to warrant a specific look. Quality-based narrative (QBN) imagines your fiction as a vast array of storylets -- little chunks of story that might have choices within -- often presented as cards. A player might select a particular card, which might be interpreted as exploring a place or talking to a person, and the scene will play out.

The "quality" mentioned in QBN is that a player will often have a wide array of qualities, which can encapsulate all sorts of things like player character attributes, money, inventory, factions, temporary afflictions, story hooks, progress in a chase... Cards might be presented or withheld from the player because of their current qualities.

Parser games tend to have an emergent story, and choice games can provide a very explicit story. QBNs are somewhere in between. Cards will have an explicit narrative, but in-between cards provides something more implicit. You might play a sequence of cards and fail at all of them. They aren't particularly linked, but your narrative as a player might be "Today my character is having a rough go at it."

The primary example of QBNs is Fallen London who developed and perfected this form. Others exist like cyberpunkdreams, but none have had the same success.

Interactive fiction continues to grow and evolve. It's a niche form of gaming, but is often a frontrunner in experimenting with different approaches to interactivity and storytelling. Many of the tools in this space are free to use. There is a small but vibrant community of developers and players.

• Interactive Fiction FAQ by the IF Technology Foundation.

• What is IF? at IF Wiki.