• Home
  • Blog
  • Gallery
  • Pricing
  • Home
  • Blog
  • Gallery
  • Pricing
Start creating

Create.Play.

Creator blog

Home/Blog/Interactive narrative

Make Text Adventure Objects Operable: Which Usage Clues Should Descriptions Reveal?

The player is stuck in the "old greenhouse" and enters "unscrew," "turn," "open," and "twist open" at a brass valve, but the system only replies, "You can't do that." The problem is usually not the verb library, but that the description does not provide usage clues. An operable object description must simultaneously reveal four types of information: material and structure, position and orientation, current state, and connections. Based on these, players infer "what can be done to it," rather than guessing the one unique verb in the author's mind.

D
DramaFork Editorial TeamInteractive storytelling and AI production
2026.09.11Estimated reading time: 13 min
A valve, fixed railing, and wooden box in a machine room show different executable actions.
Article contents
Creator blog
  1. 01Introduction
  2. 02First Look at a Description with No Clues
  3. 03Comparison After Adding the Four Types of Clues
  4. 04Write the Clues as a Range of Attempts
  5. 05Object Action Word List
  6. 06Distinguish Between Cannot Be Performed and Destruction That Can Be Performed
  7. 07Completion Check
Back to article top

Introduction

The player is stuck in the "old greenhouse" and enters "unscrew," "turn," "open," and "twist open" at a brass valve, but the system only replies, "You can't do that." The problem is usually not the verb library, but that the description does not provide usage clues. An operable object description must simultaneously reveal four types of information: material and structure, position and orientation, current state, and connections. Based on these, players infer "what can be done to it," rather than guessing the one unique verb in the author's mind.

The fictional teaching example "old greenhouse" is used throughout below. The characters, objects, and values in this example are all writing examples, not tested data.

First Look at a Description with No Clues

There is a valve in the corner of the greenhouse. Beside it on the ground is a strip of cloth, and on the shelf is a glass bottle.

This text can be read, but it cannot be operated. The player does not know what the valve is connected to, whether the cloth is dry or wet, or whether the bottle has a lid. Any verb can only be guessed. If the author writes the check to accept only "twist open the valve," then the player's inputs of "twist," "turn," and "wrench" will all be rejected. The frustration comes from missing items in the description, not from the player not being smart.

Comparison After Adding the Four Types of Clues

Clue type Static wording Operable wording
Material and structure There is a valve The valve body is brass, a hex nut is stuck at the pipe opening, and the handle points up
Position and orientation Beside it is a strip of cloth The cloth is draped over the iron railing directly below the valve, with one end hanging into the standing water
Current state On the shelf is a glass bottle The glass bottle lies on its side on the middle shelf, the mouth is plugged with a cork, and the body has a fine crack
Connections The valve is connected to a pipe The upper end of the valve connects to a rusted vertical pipe, and the lower end connects to a short horizontal pipe; the horizontal pipe leads to a nozzle above the glass bottle

With all four types of clues present, the player can infer: the valve handle can be turned; the cloth can absorb water or be wrapped around the handle; the cork can be pulled out; the nozzle is aimed at the bottle. The range of actions that can be attempted therefore becomes explainable, rather than an infinite verb list.

Write the Clues as a Range of Attempts

After the description, it is recommended to close with a sentence giving the "range of attempts," clarifying which actions are within the check and which are only environmental description:

You can turn the handle, wrap the cloth around the handle and then turn it, pull out the cork, and move the bottle out from under the nozzle. The valve is fixed to the pipework and cannot be moved away as a whole.

The purpose of this sentence is to draw boundaries. "Move the valve away" is excluded, so the player will not waste turns; "wrap the cloth around it and then turn" is included, showing that the combined action is feasible. The author does not need to list all synonymous verbs, only to explain the physical range of the action.

Object Action Word List

The same object will accept different actions in different states. First make a small table for each of the three objects in the old greenhouse.

Brass valve

State Actions that can be attempted Not suitable to accept
Handle points up, nut is stuck Turn the handle, wrench the handle, wrap cloth around it and then turn Pull the valve off directly
Handle has been turned halfway Continue turning, turn back to the original position No removable structure is described, so the handle cannot be taken off directly

Wet cloth strip

State Actions that can be attempted Not suitable to accept
One end is soaked in water Pick up, wring out water, wrap around the handle, wipe Light on fire (the environment gives no fire source)
Wrapped around the valve handle Pull tight, untie, turn together with the handle "Use" alone

Cracked glass bottle

State Actions that can be attempted Not suitable to accept
Lying on its side, corked Stand upright, pull the cork, move away; smashing requires confirming the consequences of destruction Cannot collect water without pulling the cork
Has collected water Pour out, put back on the shelf, place under the nozzle Check whether it is already full before continuing to add water

The word list is not a whitelist for checks, but a self-check tool for the author: for the actions listed under each state, corresponding clues should be findable in the description.

Distinguish Between Cannot Be Performed and Destruction That Can Be Performed

When an action cannot be performed, point out the missing condition; when it can be performed but will damage the object, first warn of the known consequence and let the player confirm. Do not write both types as "no." Synonymous verbs should be understood by intent, and everyday wording should not be treated as an error.

  • Input "light the cloth strip": the cloth strip is wet, and there is no fire source in the greenhouse. The feedback can say, "The cloth strip is still dripping, and there is nothing around that can start a fire."
  • Input "smash the glass bottle": first respond, "This will lose the water-collecting container. Do you still want to smash it?" Only after confirmation describe the fragments and change the object state to damaged.
  • Input "open the valve": understand and execute it as turning the handle; only if the player explicitly asks to lift off a lid that does not exist should you explain that the structure does not support it. If the words differ but the action intent is the same, the same handling should be used.

Include clues in the feedback so the player can correct the action next time.

Completion Check

After writing a passage of object description, check it against these five items:

  1. Do the four types of clues each appear at least once?
  2. Is the "range of attempts" written out and is one obvious but invalid action excluded?
  3. Does each state have an action word list, and can the basis for every action in the table be found in the description?
  4. Does misuse feedback point out the missing clue, rather than only giving a negation?
  5. Is it avoided that only one unique verb is accepted?

If item 5 does not pass, change the check to accept synonymous actions or combined actions, then go back to the description and add clues.

When AI-generated narrative is involved, these clues must be written into the prompt constraints; prompt constraints do not guarantee that every output is correct, and the author still needs to manually check whether the generated result preserves material, position, state, and connections.

Next step: pick one object in your hands that is blocking the player, add the four types of clues according to the table above, and then write one sentence giving the range of attempts.

Explore product capabilities Play interactive stories

Updated 2026.10.04

Explore this topicInteractive storytellingDesigning text games
PLAY / DISCOVER

Related works and experiences

Browse the works index
Character IP

Spirit of the Other Shore Flower

She reads the threads of old debts and never lets one go unpaid.

Character IP

Pip

Wants to be reliable, but keeps mistaking 'looks like' for 'is'.

Character IP

Lucian Vale

Prophecy rejected you. His office still has openings.

Character IP

Inverse-Fate Sword Embryo

He copies fate records to prove they lie.

Continue reading

Browse more articles
A courier sorts three sealed letters across a miniature coastal terrain with tide pools and a hill path, with route choices…
Interactive narrative2026.09.15 · 14 min

How to Turn a Letter-Delivery Premise into a Text Adventure: Make Route, Deadline, and Recipient Into Choices

To turn a letter-delivery premise into a playable text adventure, the core is splitting “who the letter reaches, when it arrives, and what you gain along the way” into three mutually constraining quantities.

Three letters, a repaired lamp, and a returned small boat correspond to the promises the ending needs to account for.
Interactive narrative2026.09.14 · 16 min

How Text Adventure Endings Should Respond to Opening Quests: Use a Promise Payoff Table to Check for Omissions

The ending should make readers feel that "this matter has been accounted for." The way to do this is to list the promises planted at the beginning one by one, and at the ending state clearly whether each one is completed, partially completed, abandoned, or intentionally left open. Below, a promise payoff table is used to walk through the entire process, followed by three fictional ending comparisons.

A fictional flooded town's radio desk separates two message reels with their own source tokens, while through the window a…
Interactive narrative2026.09.18 · 20 min

How Can a Post-Disaster Town Radio Story Make Choices Land in Visible Consequences?

Treat radio as a clue that walks: it starts from a source, passes through the host's confirmation threshold, the transmitter's coverage range, and the choice of broadcast timing, then lands on different listeners and changes their next actions. The Riverbend Night Radio is a fictional teaching example: on the seventh day after the flood receded, the town's Riverbend district is split between an uphill shelter and a downhill old neighborhood, and the night radio is the only stable shared channel between the two.

From a story to a playable world.

Create interactive narrative games with DramaFork: connect scripts, characters, shots, and choices into a playable first version.

Explore & play

  • Explore gallery
  • Interactive film
  • Text game
  • Character IP
  • Rankings
  • Works index

Create & collaborate

  • My creations
  • Film studio
  • Text-game studio
  • Character studio
  • Creator partnership

Reading room

  • Creator blog
  • Story topics
  • Article archive

Story guides

  • Interactive storytelling
  • Making interactive films
  • Designing text games
  • Character and worldbuilding

Plans & credits

  • Pricing
  • Credits

Policies & support

  • Privacy
  • Terms
  • Cookies
© 2026 DramaFork/DramaFork · Interactive narrative games
Press Enter to send, or drag away and release.