Today, we were inspired by the group party game, where players each contribute a sentence to the story without knowledge of any of the previous lines. So we took an idea, a spun a twist into it.
What it does
So today we bring to you, Too Many Authors, a game where players all work together to write one story. However, they face two big challenges: one, the players don’t get to see the sentences written by the players before them. Instead, they see images generated by those sentences, images generated using Google’s Natural Language Processing machine, and an image search API, with which the authors must guess the context of the sentence they need to contribute to the story. The other big challenge is that with any story, more than one author is Too Many Authors.
So, when you enter the game, you have three choices. One, you create a new game, generating a random URL which will act has the code for that game. The game will have three rounds, in which one round is completed after ten sentences. Authors can join and leave the game when they wish, but when the specified number of rounds is reached, the game concludes, and the last author, must look at the body of text selected and choose a title.
Two, the authors enter a game code, and enter an ongoing game. Three, the user can view archived stories.
How we built it
We built the project using reactjs, html, and css. In order to save user outputs into a database, and quickly retrieve the data back and forth, we implemented Google Firebase, and for the Natural Language Processing, we used Google’s Cloud Language processing, and for the advanced image search used an API by Qwanta.
Challenges we ran into
Sourcing images was a primary challenge we encountered, switching through multiple image sourcing platforms, in a search for one that was both accurate, and somewhat simple to implement. Finally, we needed up using Qwanta, a French search engine. Other than that, we ran other problems, while implementing Firebase and the Cloud Natural Language processing machine.
We made a fun and engaging game, that everyone of all ages can enjoy. Its game where team of family, friends, or just fellow writes come together to create a funny, tragic or just plain mundane story under pressing circumstances, where they need to both interpret the images before them, and deal with the fact that there are just Too Many Authors.
What we leaned
We learned a variety of skills with this project, mainly in using third party services within our application. Our members learned how to implement Google Firebase and the Cloud Language processor efficiently, and about the existence of easy-to-use French search API, Qwanta.
Also, we totally exposed the flaws in the current tier of natural language processing, but let’s not focus on that.
The ideal future for this type of game is for it to develop into an interesting type of sport, where the open and unsure of the game can be used to create “masterpieces” or works that require a great deal of creativity and foresight from the authors involved. In the meanwhile, it’s a fun and engaging activity that can be played with your family and friends, guaranteed to make any gathering more fun than it already is. Or just fun, I don’t how interesting your families are.