DEV Community: Fran Tufro

interactive flashbacks

Fran Tufro — Tue, 14 May 2024 11:10:39 +0000

In traditional writing, flashbacks serve as a tool to provide crucial information about the story that the reader wouldn’t otherwise know without delving into past events.

As a narrative device, flashbacks retain this function in interactive storytelling, yet their interactivity also allows us to shape and influence future events .

In Dance of The Spirits , we introduce the player to April, the protagonist, through a flashback where they select certain characteristics.

These characteristics subsequently influence the available options and even the reactions of other characters as the story unfolds.

I believe that beyond their informative role, flashbacks offer valuable opportunities to “configure” the game through player choices that impact the main storyline.

What are your thoughts?

sunday prompt #1 - memory swap

Fran Tufro — Sun, 12 May 2024 12:47:47 +0000

One goal of this list is to assist traditional writers in crafting interactive narratives.

Taking advantage of the serene Sundays, I’d like to propose prompts specifically designed for this purpose.

If my creativity permits, starting today, I’ll post a prompt every Sunday.

If you’re interested, let me know; and if you find it boring, feel free to say so :)

You’re free to interpret this prompt however you like.

Create an interactive short story, a visual novel, or anything else you can think of.

All I ask is that if you use the prompt, share your work with me. Whether it’s finished or not doesn’t matter; I’d love to know if this is helpful.

Memory Swap : The protagonist has the ability to swap memories with others. Each exchange alters what they know, affecting their relationships and options in future interactions.

There you go.

Now, get writing! :)

probabilistic foreshadowing

Fran Tufro — Fri, 10 May 2024 12:39:05 +0000

Something that particularly interests me about the foreshadowing technique is that it is a local element connected to the global story.

By “local element,” I mean that it is something that appears in the story momentarily with the aim of providing a hint toward an important event.

As it is something that “predicts” the future of the story, it is very difficult to do traditional foreshadowing in a game where the player’s decisions matter and significantly affect the story.

In the latest iterations I made in Dance Of The Spirits, I have been experimenting with something that I will call probabilistic foreshadowing .

I am going to heavily rely on the concept ofsilent evidencefor this.

If we consider that the player makes decisions with incomplete information, we can have certain moments in the game where we plant a random foreshadowing.

Imagine a situation where you are talking to many people and they all share their mood:

<span style="color: #000000;font-weight: bold">JULES:</span>I'm feeling great to be honest.<span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #000000;font-weight: bold">LINA:</span>I'm hella tired, it's been a nightmare week.<span style="color: #a61717;background-color: #e3d2d2"></span>Lina looks over your shoulder.
You get curious and turn around.<span style="color: #a61717;background-color: #e3d2d2"></span>Bambie is approaching the Circle of Trust. 
They're wearing a dress that looks like it was made by the devil himself.
You can't quite place the color. It looks like some shade of blue.
But a blue that's been chosen for the end of times.<span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #000000;font-weight: bold">BAMBIE:</span>Sorry I'm a bit late. You were already expressing how you feel?<span style="color: #a61717;background-color: #e3d2d2"></span>Lina rolls her eyes.<span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #000000;font-weight: bold">[bambie_foreshadow]</span><span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #009999">(50%)</span><span style="color: #000000;font-weight: bold">BAMBIE:</span>I'm...<span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #000000;font-weight: bold">set</span><span style="color: #008080">bambie_feels</span><span style="color: #990000;font-weight: bold">happy</span><span style="color: #a61717;background-color: #e3d2d2"></span>Bambie looks straight at you. 
    They smile.<span style="color: #000000;font-weight: bold">BAMBIE:</span>... pretty happy.<span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #009999">(50%)</span><span style="color: #000000;font-weight: bold">BAMBIE:</span>I'm...<span style="color: #a61717;background-color: #e3d2d2"></span><span style="color: #000000;font-weight: bold">set</span><span style="color: #008080">bambie_feels</span><span style="color: #990000;font-weight: bold">annoyed</span><span style="color: #a61717;background-color: #e3d2d2"></span>Bambie looks at Lina, then at you.
    They sigh.<span style="color: #000000;font-weight: bold">BAMBIE:</span>... annoyed.<span style="color: #a61717;background-color: #e3d2d2"></span>

We use this random selection to choose silent evidence:

If Bambie is happy, it is because they want you there, and they will help you at some critical moment in the story.

If Bambie is annoyed, it is because your presence there goes against some goal they have, and therefore, later on, they will stand in your way.

The player doesn’t know this; that’s why I say it’s silent evidence. But above all, what’s interesting is that it is defined through foreshadowing.

The interesting thing about this is that it creates a space where the final outcome of the story is not 100% defined by the player and not 100% linear either.

The combination of foreshadowing and their corresponding silent evidence, mixed with the player’s decisions, generates a situation where the player’s choices are not “deterministic” but they do make sense, and that sense was developed in the game with the foreshadowing and subsequent reactions to the state as the silent evidence dictates.

PS: In the code snippet I foreshadowed thatBambie Thugis my favorite Eurovision act this year. If you noticed that, you’re my favorite subscriber, period.

natural selection of translations

Fran Tufro — Tue, 07 May 2024 13:38:37 +0000

A third consideration (because we appreciate third positions in this list) is to emulate the book publishing industry.

We can regard the “number of languages the game was translated into” as a metric of success.

Thus, the game is translated only when there is specific demand in a language that justifies the cost.

Internally, I’m torn between translations as a metric of success and universalizing access to my games.

On one hand, I favor measuring success based on demand, but on the other hand, the scarcity of supply can restrict demand for our niche games.

This poses a dilemma for which I’m unsure of a solution.

Independent games typically delegate translation responsibilities to publishers.

Unfortunately, publishers willing to secure rights for specific geographical regions are becoming increasingly scarce.

While there are numerous publishers cross-releasing games between Eastern and Western markets, finding one interested solely in publishing an English-developed game in Germany, exclusively for the German market, is challenging.

This used to be the norm, but with the advent of digital downloads via platforms like Steam, this practice disappeared.

Unlike the film industry, which still exhibits regionalism despite digital platform emergence, we lack such distinctions.

Thus, an intermediate space exists that we must navigate, with each individual determining their stance.

I plan to adopt the following strategy for Dance of The Spirits:

Translate the Steam page into as many languages as feasible, engaging localizers for the three or four most important languages and utilizing LLMs for the remainder.
Assess wishlist distribution and launch with human translations for English and any language with a similar wishlist count.
If the game generates sufficient revenue to cover localization costs (at least 10% of earnings), I’ll localize it for the next language with a substantial wishlist count.

One question I have is whether players would appreciate or object to LLM translations if clearly labeled.

I believe it’s misleading to release a game with automatic translations without clarification.

However, if the language selection features an “AI” tag (I’m not contesting marketing strategies here), would it be as significant? What are your thoughts?

NOTE :

The language order may drastically change post-launch.

In my experience, wishlists often surge, altering language distribution compared to pre-launch.

Moreover, a prominent local streamer playing your game can significantly influence your distribution.

Thus, for point 3, periodic reevaluation of wishlist distribution is necessary to determine the next localization target, particularly after wishlist spikes.

show me the numbers!

Fran Tufro — Mon, 06 May 2024 13:22:59 +0000

In the previous approach, we neglected the quality of the translation.

However, I doubt that this mindset reflects the majority.

During one of our weekly meetings,Seba Gioseffimentioned an alternative approach.

Seba creditedChris Zukowskiwith this insight, although we couldn’t locate a corresponding blog post; it might have been a tweet or similar.

What we did discover was a blog post where he discusses theimpact of translating your Steam pageand acomment on Steamworks:

Tip: If you translate the content on your store page into languages that you are considering supporting, you can look at regional wishlists for your game to get a sense of where your game might be popular and which languages might warrant higher priority for translation of your game.

This presents an intriguing approach to addressing the cost of localization with limited resources:

Translate the Steam page using LLMs.
Analyze wishlist distribution.
Conduct human translations for regions with significant wishlists and LLM translations for others (or omit translation entirely for those regions).

I place more trust in an LLM’s ability to translate marketing text tailored for Steam than in localizing an entire game accurately.

What I find compelling about this approach is its utility even when lacking funds to translate into languages with the most wishlists.

Rather than scrambling to find collaborators for numerous languages, you can focus your efforts on languages with the highest wishlist counts.

Based on my experience, wishlist numbers decrease rapidly, with only 2 or 3 languages having significant wishlists while others have very few.

Tomorrow, we will explore the third approach, in which we rely on “old” mindsets.

i don't care, i love it

Fran Tufro — Sun, 05 May 2024 12:07:27 +0000

The first approach (of three) I want to talk about is using LLMs to translate to as many languages as possible without worrying too much about the final result.

Given that LLMs have proven to be good enough in the languages I can evaluate, I am going to assume that they will work similarly in others.

Obviously, this is not scientifically correct.

It is known that LLMs have greater abilities with languages that have a lot of text available on the internet, especially given the significant distance between English and other languages.

I do not have enough data, but from what I see, I tend to trust that to some extent the Pareto principle holds true.

80/20 is almost the same as 70/30.

As for the cost of translation, it is a bit difficult to calculate, but possible.

The tokenization process (transforming the text into “numbers” so that the LLM can interpret it) varies depending on the provider and the language of origin.

For OpenAI, English is cheaper than Spanish for example:

English: 1 word ~ 1.3 tokens
Spanish: 1 word ~ 2 tokens

Depending on the model we use and the number of tokens, we can calculate the costs.

I can already say that for an interactive fiction game, the costs are one or two orders of magnitude lower when translating with LLMs than with humans.

I do not want to make ethical judgments on this [yet], I am just presenting the data.

I will not give my personal opinion until the end of the series, for now, I am just presenting strategies.

As I said in the previous email: the result needs to be playtested, unless you really don’t care and love it.

PS: In one way or another, I managed to bring Charli XCX into this 😅.

translation is not localization

Fran Tufro — Sun, 05 May 2024 00:43:12 +0000

Today I want to clarify the nomenclature a bit.

A dear reader of the list,Andreas Lopez, responded to yesterday’s email:

My wife is a linguist and fluent in 4 languages (Spanish, English, French, and Japanese). She is actually thoroughly disappointed in the LLM translations because too often they ignore cultural relevance and dialects and mix and match them which can lead to confusion for the people you are making it - the localized versions.

Especially in interactive fiction it’s not about actual translation but localizing it. We have many ‘odd’ phrases in Germany we exchange every day but a different phrase reveals your locale or homestate.

Andreas refers to the fact that translating is not localizing .

There is a difference between putting a text in another language and capturing the essence of that text to reproduce it with local forms and manners.

This is a hard problem to solve, and most game translations suffer from this issue at different levels, even when done by humans.

I will “ignore” this problem for now, because I think this is a problem that LLMs will solve eventually.

I may be wrong, because I’m not aware of the computational complexity of this problem, but I think it might get solved with future LLM generations.

I would keep my hands off of LLM just because the risk of pissing people off with bad writing is just too big for my personal pride and risk levels.

As an indie with $0 budget you might be better off asking for volunteers or just accept you might piss people off due to insensitive translations which you won’t be able to check unless you know the language you have the machine translate into.

Yes. But also maybe not?

You can’t release a game translated with LLMs just like that. That’s a given.

Getting volunteers is not always feasible, especially with games with high wordcount.

The readers of the list can guess my approach: playtesting.

You should never release anything without it being tested, and translations are no exception.

In the next emails, we will discuss some approaches for localization using LLMs (or not).

the unavoidable cost of interactive fiction

Fran Tufro — Fri, 03 May 2024 12:35:45 +0000

Interactive fiction is less costly to develop than practically any game that requires a lot of graphics, especially with real-time motion and 3D graphics.

Except in one area of development: translations.

The cost of translation is usually per word, and this makes it prohibitive to translate the game into many languages. There are games that have a small menu, and maybe some text here and there.

If I remember correctly, Nubarrón had a total of 2000 words.

Dance of the Spirits is not finished and already has 20,000 words.

Being in the era of LLMs, this might not be a problem anymore.

I have the ability to evaluate the quality of translation of an LLM in English, Portuguese, Italian, and Spanish; and the truth is that I believe they do a good job.

At least they comply with the Pareto principle, which is already better than many translators, especially those that I could afford as an independent developer.

I am not going to spend time discussing the ethical issue, becauseI already gave my opinion on it.

In the next emails, I want to share with you some ideas I have on the subject. I was going to include everything in this email, but it turned out to be very long, so I decided to break it down into three more emails to elaborate a little more on each of the three ideas I considered.

llm storytelling in 2024

Fran Tufro — Thu, 02 May 2024 11:48:38 +0000

Every once in a while I like to do a review of the state of interactive narrative using LLMs.

In the past few months there has been some progress, but I’m still not sure if we are at a point where we can create unforgettable experiences .

I wonder if this is a natural limitation of LLMs or if we still haven’t found the right way to exploit them.

For those who don’t know how LLMs work: given a series of words (tokens) they determine which is the next word (or token) most likely to make sense, also adding some randomness in between.

My doubt is if the “most likely” approach is so core to the system that the random part will never be able to write very interesting things.

I don’t mean to say that we will never be able to automatically generate interesting interactive narrative, because eventually we will be able to, my question is if the architecture of LLMs is sufficient to achieve this or if we need a different architecture.

It’s not clear that the architecture can generate complex models of reasoning that can surprise us in fundamental ways.

Until we see a new generation of LLMs that can understand essential things like time and space, keep in memory everything that has happened in our story, and maintain coherence, have a sensitivity of our mental models to manage expectations, and other skills that writers use daily, we are still far away.

PS: as a person who likes precise definitions, it’s not surprising that I don’t use the term ‘AI’. In this case I use the more correct term LLM (Large Language Model) which represents deep learning models developed to write text.

second person

Fran Tufro — Wed, 01 May 2024 12:51:44 +0000

There is a maxim when one starts writing: do not use the second person.

I have heard this many times, and it is clear that there is a literary tradition that favors the first and third person.

However, interactive narrative as we know it today has as its great-grandmother the collective narrative of tabletop role-playing games, where the second person is mainly used.

You are in this situation; what do you do about it?

I’m not saying that all narrative games have to be in the second person.

But it is an immersion tool that the traditional writer often ignores.

You have a new tool,

What are you going to do with it?

PS: It’s the first of May, Labor Day, so I sat down to write a bit. In recent months, my ADHD has gotten the best of me, and I committed to many very big things. Now I’m trying to distribute my efforts a little better. Hope I can handle it 🐌. I created anow pageon my blog and a reminder to update it to make sure I follow through.

character vs. player

Fran Tufro — Thu, 14 Mar 2024 11:06:10 +0000

In the lines whereI talked about emotions yesterday, there is a contrasting idea that I wanted to share with you.

In one of our narrative meetings with Nico Saraintaris and Sebas Gioseffi, Nico mentioned that there is a character in hisPixel Pulpfor which he receives a lot of hate.

There is a very strong phenomenon of identification in video games that, although it may appear in some books, is very rare and I wouldn't know for certain if it happens in the same way.

Players, having agency in the game, do not create a barrier between the protagonist and themselves .

This is reinforced by the concept of avatars, digital versions of oneself.

This identification, or lack of separation, generates a very strong negative feeling when the character in the game acts in ways that clash with the player's intentions.

It is enough for the character one is controlling to say something one disagrees with for a very strong negative emotion to arise.

This bothers the traditional writer.

And rightfully so, this greatly limits the type of character one can develop.

Especially in a time where cancel culture is so strong.

In a way, it makes us feel like we are obligated to create characters that reflect the opinions of the majority.

Or if they don't, that they are cute and sympathetic "underdogs".

We can also ignore this and see what happens.

Nico did that, and he received backlash from some players.

The feeling it leaves behind is not nice at all, but it is clearly a not very good consequence of interactivity.

Have you felt this? Do you have a go-to way to approach this problem?

hello, emotions? where are you?

Fran Tufro — Wed, 13 Mar 2024 10:55:12 +0000

In a way, those of us creating interactive narratives are neophiles.

We see interactivity as something positive, as a step in the evolution of storytelling.

But I also believe that, like with any technology, we have to spend some time thinking about the negative consequences.

I have the following uncertainty:

Adding interactivity and choices to traditional narratives, does it take away emotional involvement in some way?

What I know is that "gamey" systems erode deepness in some emotions, but help others.

My question mainly comes from whether in the player's mind, knowing that they can "go back" and change their decisions removes the permanence of the protagonist's decisions in a story, thereby taking away [part of] the player's emotional involvement.

I don't know if this is true or not, and I think only time will tell.

But, assuming this is true, it would mean that interactive stories are inferior to linear stories in terms of generating emotions.

What do you think about that? Is there anything that makes you think this is the case? Or something that makes you think the opposite?