Welcome to the first devlog of A cup of eternity, an O2A2 2025 jam visual novel! In these, I plan to cover each week a different aspect of developing for this game, to keep a trace for posterity. These will be separated into several parts:

• Devlog#1: Writing & origin story
• Devlog#2: Char design
• Devlog#3: Staging and background
• Devlog#4: GUI and music
• Devlog#5: Final hours

I am also pleased to announce that we will also do an extended version of A cup of eternity for SuNoFes 2025 jam (September 2nd), with planned VA and translation possibly, and other improvements in the works. So please stay tuned and follow, rate and comment on our game if you liked it! 

1/The sandwich that started it all

A few weeks ago, I was eating a sandwich at work and scrolling through the O2A2 page, reading the rules and wondering to myself what kind of plot could I possibly do if I ever decided to take part in it. It had to be something self-contained, dynamic and interesting enough even with all the assets limitations. 

There were 3 or 4 silly ideas swimming in my head and among them was the idea of a poisoning between husband and wife in a chinese historical setting where you’d have to try to survive. It had a lot of deaths, twists, and most importantly, it could be contained in one scene while still having some strategic camera movements. (This was literally what you could call food for thought. I guess the sandwich was really good that day.)

However, at the time, I had just gotten out of crunching for Josei jam where I drew backgrounds for The Man of my Dreams and nearly died in the process, and I had promised myself to rest before starting a new project. So I quickly dismissed the idea and let it be. 

However, life has a way to make you eat back your words as you’ll see in the next part.

Some of the backgrounds I did for The Man of my Dreams

2/Making a VN on impulse

Some time later, I was discussing with a friend who wanted to join a team for O2A2. So on impulse, I told them, “You know what? We could just make one ourselves!”

“Actually, I already have one idea that goes like this…”

Thus, that is how our team of two was born, even though I had initially planned to rest…

But now that it got decided, it was too late to backtrack now. Recruitment time!

3/Assembling a team and the marketing pitch to catch them all

For the first time in my life, I ended up putting up a recruitment post and joining the Meet & Greet of the jam to pitch our story idea and fill the missing spots in our team. We even made slides but then realized that nobody else did, so we hid them back again sneakily after the first try…

Here’s what the initial moodboard we presented at the time looked like. As you can see, it was a mashup of several cdrama references with lots of sneaky glances. 

Here’s how it went:

• Character designer: applied wholeheartedly as moral support and took a wild shot at applying for character designer, thinking I wouldn’t accept them in. Little did they know, I not only took them in, but really made them work :)
• GUI artist: Since they said in their blurb that they liked making complex GUI, I specifically said during my pitch that coding complexity was no problem for GUI because I was already aiming to kidnap them at the time. Famous last words as we’ll see in the next devlogs…
• Sprite artist: the one I least expected to apply and got even more pressure to make a good story because of how godly their art was… 
• Composer: this one took a bit longer than other roles because music is so subjective that it was hard for me, as someone who doesn’t know the least bit about it, to decide by listening to the samples sent to me. 

And with that out of the way, we could finally start working on it!

4/ Outline and plot structure

The branching graph with outcomes that I shared to our writer:

The aim with the first drinking event method with 2 consecutive choices that you have to choose correctly, was to weed out those first-time players. 

In that way, they only had 25% chance of making it alive after this first drinking segment by pure luck.

Additionally, bridal wine is usually drank with cross-linked arms, so that’s the method Fujun kept in mind when poisoning the cups. So to survive, MC had to surprise him by acting out of his expectations and drink in another manner (who knows, maybe in their country the custom for drinking is different but Fujun didn’t know about that…). 

If MC survives the first choice, Fujun uses a random pretext to avoid drinking then adapts and switches to MC’s drinking method next. Once again, MC has to act unexpectedly and try the other method.

So in a way, those who are a bit familiar with that wedding custom and use their logical thinking could actually survive on the first try.

In the next conversation segment, I chose to add several layers of uncertainty at once: what to eat, what to feed, what topic to talk about first (the order is important!) and whether or not to poison or switch the cups while not getting caught if you do (and how many times).

That quickly added complexity to the whole branching for outcomes so we had to stay methodical and make a tab to count the affection points for the final drinking event, if MC managed to make it through the eat & talk segment alive…

5/ Interactivity and making meaningful choices

Here’s how our tab for deciding the final outcome after 2nd convo segment looked like:

Endings outcomes depending on affection points and poisoning state

The actual logic is even more complicated than that in the final code since it also takes into account the number of times the cup got switched too to stay logical…

But for the sake of simplicity, we’ll stop at this points tab.

As you can see, the max amount of points a player could get in a single run is 4. They start with 0 and could even go to negative numbers if they choose too many bad options.

Get to know your bae before smooching him:

We used the kiss as a multiplicator for affection points in order to punish those players who got distracted by his pretty face and chose to kiss him as the first conversation topic without trying to get him to appreciate MC first. 

So if Fujun likes MC, the kiss will add even more points. But if he already doesn’t like MC, then he will hate them even more after this.

That’s why it is important to accumulate enough points through feeding him the correct dishes and talking with him first before going for the kiss.

How the original gentle and dignified scholar ended up as a tsundere:

He originally was supposed to be a dignified, gentle but cunning and cruel groom. And he still is. However, I sent C-drama clips to my writer and one of them featured a cute acting male lead. As she was falling in love with the actor, it evolved to: what if we also add to our love interest a cute side? But he still has to stay in character…Oh we could make him flustered, an angry mess when he is embarrassed! As the writing went on, we might have had too much fun…especially with his hobbies…

Ending 4 and the petty little…!!!: 

The first version of the draft had Fujun nicely dying without making another fuss after MC successfully poisoned him. However, I felt like it lacked punch. Wouldn’t he be peeved to die in such a manner, especially after changing his mind and choosing not to poison MC because he hesitated? He would feel angry at MC but also at himself for softening up when he shouldn’t have. So the hurt and sting of unexpected betrayal in his last moments were enough to drive him to bring MC down with him. 

The line where he talks about Meng Po actually refers to the name of an old immortal granny that gives newly arrived ghosts in hell a soup with the waters of oblivion. Once they drink it, the dead forget about their past life and can now reincarnate peacefully without any memories of the past one. So it’s basically a fancy way of telling MC to go to hell with him :) 

Ending 3: the most probable one?

One other possible ending for the outcome where both poison each other could also have been mutual double kill where none survives. However, we chose to make it more interesting by having them doubt each other, so in the end, nobody trusts the other enough to drink from their own cup. 

The reason Fujun stops for a while after the first two poisoning attempts in the first drinking method segment is that he’s puzzled by how MC managed to evade his plans. Was it a coincidence, or were they more smart than he thought? So he decides to take the time to assess them a bit during the eat & talk segment before making his final decision. 

Ending 3 is the one where they both recognize the wit and cunning of the other. It can’t really be called love yet, but an ending where they end up in a partnership of sorts and reach a rational understanding of each other. 

Ending 2: the hardest to get because it’s the least likely to happen

Most players might consider it as the true ending though since it’s the one that’s the hardest one to get. You’d have to make nearly no mistakes in your choices and play a certain number of times first to know all the choice patterns and elaborate a strategy to get him to like you, while also not taking any action against him. This is actually pretty improbable, especially considering the fact that by then, MC and the player would certainly know he was planning to kill them. Hence why, if you don’t aim for it, it would be pretty hard to get that ending organically for most people. 

It was also the hardest to read and reread again while debugging and playtesting the game because I had to cringe internally when looking at the final lines but still chose to keep them, in order to fit the ending’s title.

Because the title actually came before the writing, and I wanted to make a play on words that related to the wine drinking theme…

Ending 6: the mysterious third-party

Who else tampered with the dishes? Who else would gain from killing them? That is a plot point which we wanted to expand on but didn’t have enough word count left to reveal it in the script. Maybe in a longer version of the game… 

6/ Editing for 1k words

Writing for O2A2 was quite a challenge since we wanted to have a lot of endings to explore all the directions the plot could go to its fullest and tell the most plot points possible while still staying under the 1k words limit.

So right from the get-go, it was decided to go for a dialogue-heavy script to make the plot progress as fast as possible in the least amount of words.

In the process of editing and rereading the script to detect logical loopholes, we ended up adding more and more endings and alternative lines to cover those extra cases… and each time having to cut something somewhere else in the script to accommodate the addition of the new lines…

You can see that in the number given to the endings: these were actually the order in which each ending got created. Hence why ending 2, which is one of the hardest to get, is not called ending 6 as the end route but is actually in the middle… 

There was a lot of back-and-forth where we each had our favorite one-liners which we wanted to keep in the final script. And in this process of negotiation, new ideas came to mind and the endings got more and more outrageous (cough cough ending 2)...

It was quite an arduous but very fun process to write it, but also quite frustrating because we sometimes had ideas which made us laugh and kick our feet but when reduced to their essential lines, it ended up in the script as two lines only…

Afterword

While the script was being written, we also started on other aspects such as character design, which we’ll cover in the next devlog.

Please stay tuned for the next one and thank you for playing!