Mailbag: Breaking into Writing for Voice UI

Dear Emily,

[personal information redacted] I have been following your articles for a long time and I decided to write to you because I am stuck: I would like to expand my skills in the game design field but at the same time I started to get interested in voice applications. I have read that writers and copywriters can make a great contribution to the development of VUI, so that they can understand the language and the context in which conversations take place. I saw that you managed to combine these two areas – game narrative and natural language programming – so I would like to ask you where I can start if I wanted to take this career path? What skills should I focus on? I’m looking for courses on platforms like Coursera and Udemy, but it’s not clear to me what criteria I should consider for the choice. Except for HTML, I don’t know the development software that are proposed, all I know is that I am interested in understanding how an editor, who deals with finding the best plot structure for a story or making characters believable, can contribute to the development of an Avatar or a dialogue flow, for example. And if companies are interested in this type of profile.

I definitely wouldn’t want to discourage you looking into natural language processing if you think you’re interested in it: it’s a fascinating field, and currently under lots of demand.

I didn’t approach the subject with Coursera or Udemy myself, so I don’t know the offerings there very well, but I would imagine that there are introductory courses that would explain a lot of the basic ideas and tools. Another way in would be to look through the resources listed here.

It’s also possible to play with transformer-based language models using Google colab notebooks. These models take a huge amount of skill, data, and computation time to build, but once they’ve been trained, they can be used in a range of applications. For instance, this notebook by Max Woolf will let you experiment with a trained GPT-2 model, which has applications both in generating text and in creating machine translations (among other things).

Then there are also sites such as https://huggingface.co/ where groups doing active research regularly post their progress, sample code, or trained models. You would need a good grounding in the basic concepts in order to make sense of these.

That said, you might not need all of those skills even if you were building your own voice-based system from scratch.

Continue reading “Mailbag: Breaking into Writing for Voice UI”

Mailbag: Statefulness without a Parser

Hi Emily, I appreciate the content you create to further the IF community.

I’m curious what games or platforms stand out to solve some of the issues you listed in your parser article 9 [years] ago.

I’d like to create a text heavy game with detailed world state and want to research the projects that handle these situations the best. Specifically UI driven interaction from the player. 

[I then asked for confirmation that the writer would like to see this handled as a mailbag post.]:

I’d love to see a follow up on how you feel things have changed. From the perspective of a “traditional” game developer IF seems stuck in an award state [sic] of being too gamey for readers and not enough mechanics for gamers. It’s a hard problem to solve as most people aren’t writers, programmers and designers , that’s a lot of skill sets to tackle.

I’m not quite sure what “an award state” is, but maybe “an awkward state”? I disagree, though. “Not enough mechanics for gamers” or not, games from 80 Days to Choices and Episode to the works of Choice of Games to Failbetter’s entire oeuvre are making enough money and attracting enough attention to support quite a few small to medium studios. And that doesn’t touch on the audio IF, the visual novels, the interactive film, etc., etc., etc. Interactive fiction, broadly drawn, is doing fine. And I know quite a lot of traditional game developers who think so, too.

But okay, let’s set aside that part of the question. The question is about how to do UI for a game with a lot of world state and a lot of text, but without a parser.

Continue reading “Mailbag: Statefulness without a Parser”

Mailbag: Moments of Non-Choice

This is a slightly unusual mailbag post because the question was asked in chat context, but it turned out to be something where I felt a number of other people would be interested in the answer, so I’ve paraphrased and expanded what I said there.

I have a story point where the protagonist has to do something. It feels bad not to offer them any choice here, but if they don’t do the thing, then the whole plot comes apart.

I have a bunch of tactics to offer here, depending on where in the story this is happening and what it means.

Tricks for the start of a story

What Lies Beneath the Clock Tower is partly about engagement vs withdrawal, and its start-of-adventure choice has more thematic validity than most

Start the story after the choice has been made. “The protagonist has to do this thing” is a pretty common situation at the beginning of a game where we’re looking at the inciting incident for the story. You usually don’t want to allow the player to choose not to go on the quest for Smaug’s gold. It’s better to assume that, if the player’s started a game about this quest, they want to play the quest, and we should just get on with the first interesting choice that happens after they’ve already committed.

There are very occasional cases where I think an aesthetic argument can be made for including such a choice, but they’re very much the exception.

Tricks that work later

Shift responsibility for the incident. You can decide that some other, external force is responsible for the protagonist’s bad situation — which may then require some setup to prepare. This often works fine in an action/adventure-y scenario where we expect that misadventures will regularly occur to motivate the plot.

Alternatively, we can make the bad twist into an unforeseeable but inevitable result of an action that was perfectly sensible for the protagonist to do. Perhaps they’ve rescued a puppy, in line with their characterization as a lover of animals, only this particular puppy is the carrier of a disease that sickens all the other animals in the shelter. This is where the idea of the expectation gap becomes useful.

Provide strong motivation in the choice framing. The more one digs into this, the more the difference between interactive and non-interactive story starts to melt away. A non-interactive story can force its protagonist to do something stupid on command — but the viewer still wants to understand that this act is in character.

Maybe the protagonist did steal a car and take it for a joy ride and get arrested — but they did that because it’s a flashy sports car that their asshole brother bought and drove to Thanksgiving dinner just to show off how much better he’s doing, and it pushed all of the protagonist’s buttons at once.

The thing is that if you put that same level and quality of setup into an interactive story, and then you offer the player the choice

  • Eat my turkey, keep my mouth shut
  • Join Dad in congratulating Ryan on his brilliant career choices
  • Pretend I need to pee, go outside, break into Ryan’s Porsche and drive it to Vegas

…there are pretty decent odds they will want to choose three.

And, if not, it’s also likely they’ll understand what the interface is communicating about the character if it’s greyed out the first two options as unavailable.

Dialogue in Neo Cab can be available or unavailable depending on the protagonist’s evolving mood. Disallowed choices options are conventional enough that most players will understand — even if the protagonist is actually forced to be in one specific mood at the key moment when you want to control their actions.

Skip the moment of decision. Use an act break or the space between episodes to skip ahead in the story until after the protagonist has done this.

This effect is distancing. The previous approach asks the player to think as much as possible in character with the protagonist, adopting their motivations and feelings so much that they do something that might not be in their own interests. This one, by contrast, pushes the player out of the protagonist’s head. Both can work, but they achieve different things.

You then do still need to reconcile the player to what’s just happened, though — justifying the protagonist at least in retrospect is often going to be critical to the player’s sense of themselves.

(I think my all-time least-favorite example of this is in Emily is Away, where control of the relationship is taken from the player in such a way that the protagonist engages in what could be construed as a dubiously consensual situation.)

One form of reconciliation is to make a mystery out of why the protagonist did this, and have the explanation gradually emerge through the next segment of play. I urge you not to motivate this via amnesia unless you absolutely have to, though.

Another approach is to show the aftermath of the Bad Choice, then tell what led to it in flashback. Think of all the TV shows that start with a shocking incident, then go back with “72 hours earlier…” to show how we got there. It’s a bit hackneyed, but it can work; and telling part of an interactive story in flashback mode means that we can ask some different types of question during this part of the story.

Which brings us to…

Useful any time

Offer a form of choice other than “what do you do?” Here are things you can ask the player instead:

How do you do this? Questions of method rather than intent are really common in Choice of Games works, and they feed into protagonism/identification forms of agency generally.

With what resources / at what cost / with what benefit? A bit related to “how,” but this choice gets the player engaged with the stakes of this part of the story.

Why do you do it? Motivation questions allow the player to put their own spin on an event. At the start of the game, this kind of question might let the player pick a backstory for their protagonist; later, it might have some other functions, like setting a new goal or calling back to a previous story outcome.

How do you feel about it afterward? A reflective choice with a slightly different flavor than the motivation question, often good to use as a character beat or a quiet moment between more action-y elements.

And then there are a few more esoteric options:

When, where, or with whom do you do this? All of your standard journalism questions are fair game for a choice point. And when/where/with whom questions can be good setup for an exploration or investigation sequence. Where do you go first? is a very common choice in a mystery scenario.

If there isn’t an obvious exploratory meaning, though, these questions may take a bit of framing to make them interesting — why does it matter where the player does the action? The answer to that might vary a lot depending on the narrative.

Sometimes from a narrative point of view it’s easier to think of these as resource/cost/benefit questions that just happen to be pegged to secondary characters or in-world locations.

You do this. What is the result? This one really flips the script, and sometimes it will feel deeply weird. But it can be a way to invite the player to co-authorship (at the more extreme/daring end of the spectrum). Alternatively, especially at the beginning of play, it can again let the player establish something about their protagonist’s family and home life. An example:


You put on a black leather corset with the red ribbon ties, and head for the front door. At which point…

  • Mom completely flips out — something about the fate awaiting all immodest women — but I’ve heard it before.
  • Mom completely flips out — wearing real leather is going to destroy the planet! — but I’ve heard it before.
  • Mom doesn’t look up from her laptop long enough to notice.

The corsetry event always happens, with whatever inevitable consequence, but we’ve given the player the chance to pick one of three possible conflict engines with the protagonist’s mother: she’s a workaholic too busy to give us attention, she’s conservatively controlling, or her ecology-focused activism makes her hard to live with.

Mailbag: IF and game writers

As you studied IF for a long time, would you think IF writers are videogame professional writers, or those are 2 separated groups of people ? And would you be able to estimate the number of IF writers ? 

[I’ve edited out a number of side points, but the longer letter also made clear that the writer is interesting in building a tool and/or platform, and is curious about the possible market for this tool and the ways people might use it. — Ed.]

Headcount is a very hard question to answer, because there is no longer one single “IF Community” — I mean, really there never was, but lately there has been much more of a diaspora. The Twine space or the Choice of Games space don’t always overlap and aren’t even always that visible to people working with the parser interactive fiction tradition (and vice versa); and then you’ve got huge, huge numbers of people who are doing interactive story of some kind but only within an app like Episode.

As for the skills you might find among IF writers, that’s again a spectrum. If you went back to 1998 and looked at who was writing IF then, you’d have found a community that was somewhat-to-very technically skilled, since writing code was necessary for almost all the projects that called themselves IF; but also pretty much entirely amateur, since almost no one had ever done any IF writing for pay. 

Now by contrast you would find that there are

  • gig-economy creators who have created technically simple projects and sold them on platforms (like Episodes or Kindle ebooks) that are designed for low barriers to entry; or who have built up some following on Patreon
  • technical inventors and academics who have built very complex and ambitious projects but never worked in the game industry at all; 
  • successful authors in linear media like Cassandra Khaw and Max Gladstone who have made a crossover to working with interactivity;
  • grant-funded interactive media artists who are often experimenting with form or subject matter features that wouldn’t necessarily work as for-market projects
  • well-known professional game writers and/or designers like Liz England, Meghna Jayanth, Jon Ingold, Brendan Hennessy, or Leigh Alexander; these may have background experience with interactive fiction or might still create some text-focused projects at times

So some of those people might be drawn to a new platform. Others, though, already have plenty of venues to publish, or on the other hand are engaged in IF precisely because it gives them a space where they can experiment, build weird science projects, or create personally meaningful art.

Then the question becomes: what could a new platform offer that would appeal to the largest possible subset of the above?

  • the ability easily to deliver experiences that people currently want to build but for some reason cannot — but then you have to figure out what there’s the most thirst to do — I haven’t done this in a few years so the answers are no doubt very out of date, but in the past I’ve run some informal interviews and surveys to find out what people found most vexingly absent from current platforms. Those surveys come from 2014, so there’s definitely room for new research here
  • an audience
  • money — but money typically follows from the audience, and if you had a space were new works got thousands or tens of thousands of readers, you’d definitely find at least some IF authors bringing their work there even if no cash were exchanged.

Mailbag: IF for Reinforcement Learning

Hi Emily

I’m a PhD student working with Prof. Mark Riedl at Georgia Tech and Microsoft Research Redmond. I am currently working on making AI agents (specifically using reinforcement learning) that play interactive fiction games (text-adventure games in the vein of Zork) in a non-game specific, generalizable way.

I was advised by Prof. Janet Murray that you would be the right person to help answer a question I had regarding these games, given your expertise in interactive fiction. If you have a list of such games (e.g. those given here https://github.com/microsoft/jericho#supported-games), is it possible to identify a subset of maybe ~10-15 of them that reasonably cover a majority of all interactive fiction games in terms of game structure, i.e. linearity of progression/score accumulation from the perspective of a learning agent? If it is possible, what would this set look like? Any insight at all would be great.

Nice to hear from you — I’ve been keeping an eye on this space as people have been publishing about it recently.

I’m not sure there’s a perfect answer to this, since IF is hugely varied in how it handles world model, score, pacing, etc. Also, your list here skews very much towards early interactive fiction, which means it doesn’t cover some of the formal experiments that came along later.

I also don’t remember how score works in all the games in this list — some of them I’ve not played, or played a long time ago.

However, with that in mind, here are a couple of categories that represent some fairly standard game structures:

Short or medium game in which score is given out rarely — Lost Pig (max 7)

Short or medium game in which score is given more frequently — Meteor etc. (max 30), Balances (max 51)

Long game in which score is distributed fairly frequently throughout — Adventure, Zork; possibly Enchanter and Sorcerer also; Anchorhead, as I recall

And from your list, I recall these being ones that might pose an interesting challenge:

Curses — it’s long, it’s complicated, it does have a scoring system which it doles out gradually, and it also does a trick (if I’m remembering right) where it actually at one point deducts score from the player again. 

Wishbringer — this one’s interesting because there’s a scoring system that reacts to how many times you’ve used the magic stone in the game — so the more you use wishes, the easier the game becomes, but the lower your final score.

Hunter, in Darkness — doesn’t keep score. There’s also a procedurally generated maze in this, which I would expect to make it very challenging indeed.

Thinking about games not on your list, here are some other formal extremes that might be interesting to try to reason about; all of these can be found on https://ifdb.tads.org/ and in most cases they’re available for download.

ASCII and the Argonauts — an intentionally short and simple game that gives a bunch of +1 rewards for doing basic tasks; the relatively small verb set might make it easier than some of the other games.

Aisle — a game that takes one move to play, and for which many different verbs are available; there’s also no score. It’s hard to imagine how one would use reinforcement learning on this, but it represents one extreme that might be valuable for purposes of thought experiment.

Adventurer’s Consumer Guide — as I recall this one gives out a pretty steady stream of +1 point rewards, rather than only a few or only rarer rewards, so it might be a nice counterpoint to some of the others.

Savoir-Faire — a game of mine, and I suggest it just because I happen to know it well enough to know how the rewards work; there are frequent opportunities for scoring and some rewards are bigger than others.

Bronze — a game that I wrote that keeps track of how many rooms you’ve explored and triggers certain narrative events when you’ve found more of the space, so you could use the explored-rooms count as a secondary signal to score and probably get some useful reinforcement out of that aspect as well.

Captain Verdeterre’s Plunder — gives you a score based on how much loot you managed to rescue off a sinking ship before it goes under. Genuinely an interesting optimization problem; human players have competed to try to come up with the highest-score possible traversal.

Journey to Alpha Centauri in Real Time — as the name would suggest, this takes place over a certain amount of elapsing real time and therefore it’s not possible to finish, because it’s representing a very long journey in space.

Rematch — a game in which the challenge is to figure out a single very long command that will solve the game in a single move, and in which there is a cyclical pattern to the initial world set-up. (I think this one is not a z-machine game, so it might not work with Jericho.)

Zero Sum Game — starts with a score and counts down to zero (but this may be less interesting than the others since you could just reverse the sign of the signal and wind up with something equally valid).

Hadean Lands — fiendishly hard puzzle game, in which instead of score you’re gaining access to lots of objects which could arguably be used as a proxy for progress. Also features areas where the player has to do similar things in slightly different ways.

Mailbag: Pedagogical Uses of IF in the Classroom

Pardon, may I ask for some suggestions of resources (articles, short essays) about Interactive Fiction in classrooms? Thank you very much.

And then when I asked whether they were looking for IF taught as the subject or as a means to learning other things:

Surely I’m interested in IF used as a pedagogical tool, as broad as possible (in terms of grades, subjects, case history). An introductory (and inspiring!) blog post would be very useful. Thank you!

Interactive fiction has a long history of classroom use at most levels (a little bit of elementary-level use, but then more in middle school, high school, and university teaching). Several researchers have built syllabus materials that make extensive use of IF; have published about IF-related pedagogy; or have given talks and workshops about how to teach using interactive fiction. At NarraScope, for instance, there were some workshops on this topic as well as a panel on IF and education.

I haven’t done much hands-on work with this myself, but here are some links that may be useful in this area:

Continue reading “Mailbag: Pedagogical Uses of IF in the Classroom”