After our initial hack failed, and with only 12 hours of time remaining, we decided to create a proof-of-concept that was achievable in the time remaining. As Twilio was a sponsor, we had the idea of using SMS to control a video game. We created Hackermon to demonstrate how this technology has potential, and as a proof-of-concept of more practical uses.
What it does
Controlled entirely via SMS, two players can select a knockoff Pokemon and fight each other, with the ability to block or attack. The game is turn based, and has checks to ensure the person texting the API is the correct person, so cheating is effectively impossible.
How we built it
Challenges we ran into
Sleep deprivation was a major challenge that affected us. Trying to focus on learning a new API and developing with a new framework was very challenging after being awake for 22 hours. However, having to prototype something so rapidly was very rewarding - we had to carefully prioritise and cut features in order to create a demoable product in time.
What we learned
Our initial idea for a project involved using Facebook's Instant Game API. We discovered that many of Facebook's APIs aren't as documented as we expected, and some of their post-Cambridge Analytica security features can cause major unexpected issues.
This was the first time we'd ever used the Twilio API, and it was great to learn how powerful the platform is. Initially, we'd never had to handle getting requests from the backend to the frontend in Node.js before, so managing to get this to work consistently was amazing - even though we know it's not done in the most efficient way.
What's next for Hackermon
While the game itself is only a basic proof-of-concept, the mechanic of using SMS to control a game has many applications. For example, a quiz webapp used in university classes could accept inputs via SMS rather than requiring students to download a clunky and badly designed app.