Things I learned from making my first VN

I wanted to share some learnings I had while making my first ever visual novel (and playing other people's entries) in hope they might be useful or entertaining for others.

Use Ren'Py, stupid!

There is a saying about "No one ever got fired for using xyz" with xyz being some proven technology. Well, I don't know how many people actually have jobs using Ren'Py but I can promise you, no user will ever be angry at you for using Ren'Py.

Not only does it offer great quality-of-life features that your players will love, but it is also insanely flexible. The dice rolling mechanic for my game took me just a few hours to implement, with zero prior knowledge, that is, and is super ergonomic to use in the script.

Ren'Py is especially great for jams, where people might play multiple games in a row. Being able to use the same shortcuts throughout different games makes life so much easier for players! And you get to concentrate on the content of the game instead of worrying about the technical stuff.

Of course, if you have a good reason to use another engine, more power to you! I am always glad to see some innovation! Some of my favorite games of this jam were non-Ren'Py ones.

In short, I wouldn't have been able to deliver my game without Ren'Py. 

Writing for a VN requires some adjusting.

I actually based the game on a short story concept that I had in my drawer.

It turns out VNs require a different style of writing, so I had to get used to it. Of course, some VNs manage to pull off a "bookish" writing style very well, but that was not the style I was aiming for.

This is why the first parts are very narration-heavy while later parts are very dialog-centric. I am telling myself that it kind of makes sense, as it mirrors the player taking gradually control of the character and not needing as much narration and probably something about character development but feel free to call my writing inconsistent.

Also, I decided to use a third-person point of view, which I think works well for this story, but I caught myself writing things in the second-person and having to edit them later. So future VN's of me might be written in a second-person POV. I know, weird, but I love the Choose Your Own Adventure vibe.

Grammar checking is not optional

With English being my third language, I am doubly insecure about this. I am also a total scatter-brain that makes lots of mistakes just out of pure carelessness. 

Two things helped me to cut down on issues:

• https://gitlab.com/hlieberman/renspell/
• https://quillbot.com/grammar-check

The first is a Ren'Py specific tool for checking spelling and grammar that is comfortable to use as you can check the script itself and don't have to export the text first. The second is a grammarly-alternative that is a bit nicer to use in the free-version. 

I think it is good to use multiple tools as they can find different issues and also to not rely too much on the suggestions but also to double check with native speakers.

Adding scenes and characters costs money

My biggest problem was vastly over-scoping this game. I really hated writer-me for needlessly adding new locations and characters. 

While those things did not cost me literal money, as I did not have a budget to begin with, they still ate up precious time.

In fact, even in the last few days of the jam I did not have a clear idea if I would be able to submit a working game. Somehow I made it, and I am so glad I did.

AI will not take our jobs

I have used Stable Diffusion to create the backgrounds in this game. This allowed me to open up the world. The game has 23 different backgrounds. I wouldn't just not have been able to do that the conventional way.

Still, while this worked out great in my case, I think the fear that AI might take away jobs for artists is vastly overblown.

It worked well for me because

• I only did outdoor scenes
• Had a fantasy theme that allowed for lots of leeway

And even this had several limitations. I wanted a bonfire scene for the characters camping in the woods. The results were hilariously bad.

If you don't do outdoor scenes or fantasy/horror themed games, it might be very difficult to effectively use AI generated images.

Also, making AI-based images still requires skills. My first backgrounds were significantly worse than the ones generated later on, and I ended up replacing lots of them later on.

People are really paranoid about dice-rolls and I love it

No, the rolls are really not fudged. I pinky-swear!

So you tried to cheat fate and are wondering why you still got the same dice results? If you want different results you need Ren'Py to use another random seed. The easiest way to do that is to reload your last save.

Also, high dice-rolls only mean your characters performed the attempted task well. Now the results themselves are in the hands of the gods.

I personally feel that some parts are funnier if you fail them.

The community makes us strong!

DevTalk is the most amazing, supportive, and inspiring community I have ever seen, and I feel so humbled and amazed that I get to be a part of it!

I wish I had been active in the channel before submitting the game but I felt a bit intimidated. Turns out, there was no reason to be! The community is super welcoming.

I am so grateful for all the great, amazing, talented developers that actually took the time to play my game and give feedback! Tundraflame even took the time to play my game on Twitch!

Also, playing other people's games and leaving feedback was so much fun. Literally every game I played was so much amazing.

What's next?

I plan on making a Halloween special edition that will contain additional content. I am not sure it will be enough to justify another play-through, but you can at least expect a few more scenes. I felt like the current game does not deliver much on the soldier-killing promise, so expect some violence.

The truth is though, the game scales badly.

With currently 27 decisions and dice rolls to make, the game is already moderately complex. While most of them just add flavor, it is still a huge challenge to test all possible combinations. This is especially bad as the game isn't strictly tree-structured. Many decisions also affect things much later in the game. Now add to that the non-determinism of dice rolls and yeah, it is a problem.

I am not backing down from this challenge, but realistically, development might slow down over time and it will become a long-term project. In short, this project will never die. I will slowly keep on adding stuff.

Though I do have some other VN-ideas in my head that I would love to try out. So expect me to get side-tracked.