Craft, Parser IF, and Nouns

One of many ways in which printed text (I use the term loosely, and include digital reading as well) differs from parser IF is the problem posed by nouns. In printed fiction, nouns enjoy with verbs a place of privilege in sentences. They contribute to a sense of completeness. There is something that separates the actual from the indefinite in a well-placed noun.

This Is Just to Say
William Carlos Williams

I have eaten
the plums
that were in
the icebox

and which
you were probably
for breakfast

Forgive me
they were delicious
so sweet
and so cold

What else is there to say? The plums are delicious, sweet, and cold. Perhaps some readers have wanted more, over the years, but I imagine that most of us feel satisfied with the few-but-efficacious words from which Williams’s “This Is Just to Say” is constructed.

In a work of parser IF, on the other hand, there is a craft assumption that nouns be fully implemented. The plum ought to be edible. It should, I think, have a smell. A pit, perhaps. It might it bruise if thrown.

The icebox is the real bugbear. A player will naturally want to open it. Is it a refrigerator or an actual icebox? What else is in it? Should those things be edible, too? And so forth.

There are a number of ways that an author can avoid this trap. A common one is for a narrator to speak of out-of-play matters. A refrigerator’s manufacturer, for instance, or the game world’s political realities as they apply to agriculture. An author could keep the protagonist out of kitchens and other real-world environments altogether.

Things become awkward, though, if the player realizes that they have visited a world without things. A counter-option is to use nouns liberally, but shut down interaction with them (forgive my Inform 7 construction):

instead of doing anything to the driver's license:
	say "Leave that thing alone. The picture is so bad it even embarrasses you." instead.

Will the player get frustrated after a while? Perhaps the author might, rather cynically, describe a window.

The description of the window is "Like every other house you've ever seen, this one has windows. If you open or even look in every window that you encounter, you are bound to wind up in jail.".

It seems to me that this description would violate an unspoken agreement with a parser IF player, because of course a player will attempt to open any discovered window.

Fellow players and authors, what are your thoughts on nouns in IF? It seems that the craft of printed text and the craft of IF are at odds. Authors, what tactics do you employ to avoid the multiplicative effects of nouns in descriptions? Players, what tactics seem to work for you?


One of the most useful things I’ve read was Ryan Veeder explaining how he works very hard to write descriptions that do not include extraneous items. This is both to keep the player focused, and to avoid having to spend extra time implementing “description clutter” (my term, not his.)

I limit the space available to populate each room because I don’t want them to get too crowded. A room only needs two or four thingies in it! Five, max. And I’m talking about thingies the player needs to keep track of: Stuff that should be on the edges of the player’s awareness, like the floor tile or a window, doesn’t count. Neither do nested or intentionally hidden objects, like a desk containing a drawer containing an envelope containing a lock of hair. You know what I mean.

Limiting the number of details per room makes the world a lot easier for players to hold in their minds, and it makes implementation a lot easier for the author. This approach to design could be termed “efficient” or “lazy” but it has also been called the “school of Ryan Veeder” which is kind of nice.

This kind of goes against authors’ normal impulses of writing and painting a descriptive picture, but IF has different objectives. You have to remember - if the author specifically mentions the wallpaper, that calls attention to it and the player will naturally want to know more about it. If you’ve got fifty scenic items in a room with descriptions, it makes the one important thing the player needs to worry about less obvious.

That’s not to say don’t write beautifully descriptive games that are museum pieces about exploring things in detail, but it’s good to keep in mind how you want to focus the player on any potential forward momentum in the story.


Thanks for the link. I hadn’t read it before but find myself doing similar things. Map first, manage counts, special defaults, etc. I think I picked up a lot of that lurking here.

I think five is a good number, too, and I am usually in that ballpark. Where things get weird for me is in general locations that have no plot significance. However, the places existing seem to make more sense than not existing, like this pharmacy parking lot:

  • There is probably an undifferentiated mass of cars, there.
  • There are (automatic) doors that open as the player approaches.
  • The building is red because it is made of red brick. I don’t think that calling it “red” would be enough.
  • There is the player’s own car, of course, but that work is already done.
  • There is a sign.

None of these things have anything to do with the main throughline of the game. It’s true—a goal is to move the player through the game—but what should moving through that game be like? Nearly all of them could be ignored, but doing so has implications for tone, pacing, etc. Not good or bad, mind you, just implications.

Option 1: Iceberg method. The description of the parking lot is “You stand in the Pharmacy parking lot. Its entrance is to the north.”

Option 2: Abstract. “You are outside one of thousands of PharmCo pharmacies scattered throughout the southern United States. Because of PharmCo’s predatory markups, their customers are almost exclusively insured.”

Option 3: “Realism.” “Few other cars are parked here in the PharmCo parking lot. Like every other PharmCo you’ve ever seen, its lot is L-shaped, and a sidewalk acts as a separator between the pavement and its red brick building. Two sets of automatic doors await potential customers at one corner of the structure. Behind the PharmCo and out of view is the drive-thru pickup window.”

Option 3 isn’t terrible (five things if you count the bricks). The language isn’t especially florid; it’s pretty direct, in fact. However, the nouns in the room description don’t really move the player forward OR serve as a museum piece. There aren’t going to be a lot of interesting verbs to apply either without special effort. Some attention might be paid to the automatic doors, though they are more realistic than interesting. Not all realism is interesting.

I think I am saying that there is a wide gulf between a museum piece and an iceberg description. All of Ryan Veeder’s advice is good. The prose style of the fire department example is one way to manage nouns in a description, but there are many authors who have been successful with more verbose descriptions—there are a lot of examples in Mike’s Cragne Manor let’s play, for instance. Or in its source material, Anchorhead.

What I’m really wondering from a craft perspective is this: what are some strategies for writing fulfilling—elaborative, even—prose without getting into trouble with nouns? How to talk about things without adding “thingies?” Are players conscious of these strategies? If so, do some work better than others?


Another option is to distinguish nouns that are interactable from nouns that aren’t. Maybe print the former with an underline. This is technically straightforward, but it also changes the experience for the reader, and can lead to lawnmowering.

1 Like

In the early text adventures on machines with limited memory, objects that could be interacted with were usually made clear.

This was either using a list of objects you could see shown below the location description or by displaying those you could manipulate in a different colour or typeface within the description.

This is similar to the method used in later point and click games, where the mouse pointer would change shape or colour when you hovered over an interactive item or the object itself would become highlighted.

1 Like

It’s not just nouns, verbs have a similar problem. If your room description includes, “You can cross the bridge to the east,” players will expect CROSS BRIDGE to be implemented.

In writing I’m a fan of Flaubert’s (supposed) advice to “use three senses to make a scene come alive.” It’s pretty ingrained in my writing to include at least one sensory description beyond sight and sound.

Adding a sensory emanation is another implementation problem though. Not only does it require the additional prose response for SMELL or LISTEN commands, you have the problem that emanations like odor and loud noise requires additional coding to simulate their presence. A ticking grandfather clock can be heard throughout the south wing of the mansion, or the smell of bread grows more pungent as you approach the kitchen.


Yes, very true. I’m less conscious of that but trying to be more aware.

This is great advice. I do try to have SMELL and LISTEN responses for each geographic region, though sometimes that isn’t enough (the pharmacy and its parking lot will sound quite different).

We’ve talked a little about this before, but your use of This Is Just To Say completely encapsulates my feelings on why questions like these are best seen through the lens of poetry. Each object in a poem-- usually these are few-- is used for its ability to play the theme, to give not a just a sense of place but a sense of self, of mood, of the direction of thought. So should each object in a game be used, not just as set decoration, but as proxies for the mindset of the PC, of the tone of the game. So picking those few objects and really letting them shine is what’s important, I think, like the plum and the icebox. The rest of it can be lumped into categories that do the same:

The interior of the roadside food mart is as depressing as its sagging exterior. Dusty cans and faded boxes litter the warped shelves lining the crooked aisle to the counter, where a withered and wizened old woman stares vacantly into space by a vintage cash register.

I want you to go to the counter and talk to the woman, but you’ve played a few games and may want to X those boxes and cans, so I lump them together:

X cans
Everything on these sad shelves looks ancient and forlorn: a few dented cans of cream of celery soup, a few boxes of sickly-sweet children’s cereal: the kinds of things that cause memories of childhood to rise up in your throat like bile.

If you keep insisting on being a nitpicker:

X boxes
Focussing on these things is giving you ennui.

Get to the counter and the NPC, dude. Anyway, use everything to set the stage whether it’s important or not.


How detailed you make your IF is likely to depend on whether you’re trying to create a simulation or a game. It will also depend on whether you want a small detailed world or a larger less detailed one.

The later Elder Scrolls RPGs allow players to pick up everything that isn’t nailed down, regardless of whether it is actually useful.

Unfortunately, they removed a number of game features to do so within time and budget constraints. They were also much smaller than the first two games in the series.

1 Like

Authors of both IF and conventional prose fiction both have to provide an appropriate amount of detail to achieve verisimilitude (generally; I’m not denying that there can be successful fiction that throws this to the wind). IF authors have an additional struggle, though, to provide those details while creating an appropriate impression of their importance so as not to create hundreds of red herrings that would just annoy the player. Prose authors of mystery plots have some of the same concern, but the scale is very different: the detective isn’t going to go down a rabbit hole just because they discover the oven can be opened.

Here’s a small bit I admire in @aaroniusSand-dancer

Instead of taking the tumbleweed: say "Let it blow on. Wherever it's going is bound to be better than here."

We get a solid indication that taking the tumbleweed doesn’t play a role in the story along with reinforcement of characterization, setting, and tone in a handful of words. That’s the ideal, to have the rejection also serve other story purposes. But if you just succeed at hinting that significant things are significant while clarifying that insignificant things aren’t, you’ve succeeded.


I read the first post and immediately thought “I’m going to find Ryan’s post and link it”, so I’m glad you already did and explained it so well!

Edit: I’d have to say that every noun and verb in your text is a clue to the player as to what they should do next. This is especially the case for non-native speakers; when I play German or Spanish or French parser games, a lot of times I am totally lost by weird words I never saw before (for things like ice skate blades or moldy mouldings), so I often will try commands using the words I see in the text. It works perfect for some games, but for others that have rich room descriptions (like Phoney Island, an enormous fantasy game about Trump in German) I felt helpless flailing because a dumb person like me couldn’t figure out what was important.

I do go out of my way to avoid objects with many interactions; in my latest game I purposely avoided including a fridge, sink, toilet, or shower in a house just because there are so many things a player can do with them. For things I was forced to include, like shelves and a crack behind a counter, I relied on beta testers, since any weird thing they tried I could implement (in one game a tester managed to put 4 alcoholic drinks, some tweezers and a mouse on top of a hat). That’s why I think beta testers are so vital


Lots of good advice here, but I’ll add two more thoughts (possibly worth a cent apiece? Certainly no more):

First, I really dislike the thing in some games where a location gives you a paragraph of lovely, evocative writing, and then right under it you get “You can also see a top hat, a crowbar, and a fruit crate (which is empty)” – ugh, it takes me right out of the mood, and turns what had been a nice bit of setting detail into a two-dimension backdrop. Obviously if that paragraph is showing up because the player has moved stuff around, that’s different, but when first encountering a new area and new objects, I really want to see everything integrated. With that said, I sometimes have taken inspiration from that approach by writing my location descriptions where the first paragraph is focused on the scenery and vibe, and then add a second paragraph that lists out the game-relevant objects – even though it’s not as stark as the Inform-default way of doing things, most players have sufficient muscle memory that it helps channel attention appropriately.

Second, building off of @Zed’s example, in some sub-genres there are conventions for providing non-boring output when engaging with stuff that’s not important – so that makes it easier to implement stuff, but do so shallowly without the player getting annoyed. Like, in comedy games it’s often fine to have pretty detailed environments, since each object is an opportunity for a joke, and if you have the response to interacting with something be a joke, that’s a signal that you’ve gotten what it is to get out of it and can probably move on (this means that utilitarian, non-joke-y descriptions can be a signal that an item is involved in a puzzle!)

I kinda tried to do a similar thing in my memoir game, which had lots of scenery and miscellaneous objects that you couldn’t really do much with, but most of them embedded a memory or bit of backstory when you looked at them – hopefully signaling that they weren’t important in themselves, but providing something more engaging than a default “this is boring, you shouldn’t have looked at it” response (like, in the fourth section of my game, there’s a room with a bunch of furniture that’s deeply unimportant, but looking at it cues a reflection on how there’s too much of it in too small a space, because of my mom downsizing her living situation post-divorce). There aren’t as many conventions around this sort of game as for comedy games, of course, but I think players unconsciously picked up on the vibe – it also helps that there aren’t many/any traditional object-manipulation puzzles in that game…

Oh, last bonus thought: everybody thinks the “don’t make a kitchen” thing is a joke, but seriously, if you can possibly avoid it, just don’t make a kitchen.


So you’ve said specifically, “this location has no plot significance.” There’s nothing wrong with using locations as a liminal space, or I could see this essentially serving as an “establishing shot” like a title card in a movie letting the player know the game or the next scene takes place in a pharmacy. If that’s all they need to know, why distract them (and potentially add 10 minutes to your 2 hours IFComp playtime!) if there’s no plot to be found here?

My preference given that is actually this:

This provides good authorial voice and world building without adding fifteen things to distract the player from the plot which ostensibly happens inside.

Look at the extra work in this one:

That is about nine extra things to fool with that don’t advance the plot. The question to ask is do any of these things need descriptions?

While you can write clever prose about each item like in example #2, you need to decide what important things will this teach the player and whether that’s worth 1000 extra words. The only slightly important thing you’ve added in #3 is a drive-up window, but that’s never the first thing I’d see pulling up to a Walgreens.

Something to remember: trust your audience’s imagination. Everyone’s been in a pharmacy parking lot. Is there any honest literary value you can add by describing all those things over again?

Maybe the answer to that is yes - say if this is a zombie pharmacy or a pharmacy for fish at the bottom of the ocean! Or, if it’s ordinary give it one unique detail to be memorable. But you might not want a game of all this:

“The sidewalk, like every sidewalk you’ve ever seen, runs across the front of the building to prevent customers from needing to walk on bare ground or grass. Poured cement you observe, knowingly, also noting that the surface is divided every few feet by a neat line-like division between segments so that settling and ground movement will prevent the cement segments from stressing and showing actual cracks for a couple of years.”

Another thing to consider: every description you write will invariably introduce new scenic detail nouns that you will then be compelled to implement and describe. The more detail you put into descriptions, the more digging the player will do - especially if they don’t know your game loop yet and are not wanting to miss anything important.

If there’s not something special and plot-worthy about the bricks or the automatic doors or the sidewalks, give your reader permission to fill in ordinary detail with their imagination.*

*Unless, of course, you’re Douglas Adams and can write an entire page about a drink coaster and make it hilarious.

TL;DR: There’s something to be said for realism, but in reality, I don’t notice nor examine every smidgin of every parking lot I walk through.


This is a non-traditional view, but I hope helpful. And intended so.

In the same way that mazes (left, right) have become tiresome, the same doom I predict for certain archetypal objects (doors, keys, chests, pools of still water within which lie treasure, etc).

For so long as this Mythology was held up in the player community, it was effective. I believe that now it is going away.

Human life is about action. And language reflects that. Parsing human speech should not focus on identifying named objects, but recognise our tendency to play with them as anonymous instruments.

There has been a lot of work in this area, but we ignore it. I’m thinking of Jon Searles speech acts, for example.


My main weakness is that I’m aware that I should implement smell and taste descriptions with stuff, but I actually have aphantasia with those two senses in particular, so I literally cannot comprehend how to describe anything for those senses. About a fourth of the objects I put in my games have a feel description, and I put a listening description for anything that makes sound.

I never use smell or taste in any game I’m playing, because I won’t understand the response, and likewise I probably am not likely to use them in anything I create.

If I eat a meal, I cannot remember how it tasted after 30 minutes have passed. Ever. The only way I remember which foods I like is because I explicitly note down if I am enjoying a meal while I’m eating it. After those 30 minutes, I am hopeless to form an opinion. When I’m choosing what I want to eat, I pick randomly from the list. While I am actively eating something, I cannot describe it past “neutral”, “salty”, “sour”, and “bitter”.

Similar with smells: I am unable to describe smells as I am smelling them, and less than 2 minutes after the smell has disappeared, I cannot remember a single thing about it.

When tastes and smells are described in books, I simply do not comprehend what is written, and they essentially bounce off of my brain. When meals are described on cooking shows, I simply shrug.


The problem here is not the number of “nouns” but what you can potentially do with them. Obviously, things that are, or should be, “gettable” are the trouble makers. Any background objects mentioned in the description should be added and given a bit of text. That’s not hard. In come cases, several objects can be lumped into one with synonyms and a generic “description”.

Alarm bells are things like knives in the kitchen, water/ice, matches/candles and ropes. For example, most real kitchens have knives, just don’t mention them.


So…I’m kinda leveraging the fact that the player character is both autistic and explicitly not a self-insert for the player here. Depending on the location, if you try to take an item, Marasi will say “I feel like that belongs here, though.” (I am using this excuse very rarely, though. Like for trash bins, which can technically move, but are meant to be decorative.)

If certain actions would cause a lot of entropy, and the player would have no way to clean up, Marasi says “I don’t have the tools to handle the consequences, and just leaving a mess here would be overwhelming for me.”

I have soap bottles that can be swapped between dispensers, for example. The player can only use them if there’s a sink nearby, and Marasi will just wash her hands. If you try to use one without a nearby sink, she will say something like “I really do not want soap stuck on my hands for an undetermined amount of time, thanks.”

Honestly, I am having a lot more fun making this game than previous ones, because there’s less obligation to figure out how to handle really uncomfortable or cringe-y outcomes. However, I am still trying to make as many nouns interactive as possible.

EDIT: This was supposed to be a quote reply to @jkj_yuio but apparently the quote got removed by the system, and I don’t think the reply link stuck.


I think this is important. Objects can yield thematic or psychological insights; they don’t all have to have aesthetic or mechanical utility. Scenery is a way to funnel players in an enriching way without throwing them in a barren environment.

This is an elegant way to put it. Rejections that are thematically relevant (and hopefully interesting) help the player understand that the interactable is a subset of that which is discernable, while also developing various literary qualities (tone, mood, character, etc)

This is a good point, too. I might take a risk and say that detailed environments might enhance many non-comedic games that prioritize mood or rhetorical consistency over mechanical tasks and objectives.

Oh my god, the kitchen I wrote. To make things worse, I put two puzzles in there. Never again!

Absolutely. Instead of “Realism,” I initially called #3 “Troublemaker.” Trusting your readers is always good advice; I’m glad you’ve pointed out its application here. E: I think I could do better with this.

I do a lot of this. I wonder how many verbs might apply to such objects? Sometimes I disallow anything but examining, other times I let some verbs through. It seems like it’s better to let players know right away if something isn’t going to be “useful” in a mechanical sense.

It’s a cheat, but my WIP has only one object that a player can “take” over the course of 70k words. We’ll see how that goes over!

I think skipping the self-insert thing is working out for you! Using responses to develop character can be as interesting as mechanical responses are. Or more interesting. It all depends on the writer’s goals.

I appreciate all of the thoughtful replies!


This is a remarkable way of experiencing the world, and I appreciate your description of it.


By extension, it’s OK to change the room description as the player moves through the game, perhaps exposing interactable items only once they are intended to be interacted with. This means you don’t have to reveal the car you eventually need to drive off in/steal the parking ticket from/otherwise interact with until your character has reason to want to ldo that sort of thing and isn’t just thinking, “Oh, this is the car park I must go across to get to the grocery across the road”. (You might also need to add a changing description to the car to explain why you’re not able to interact with you want it to be possible, if you’re worried about replayers short-cutting part of the game).