Joachim Breitner's Homepage

A while ago I set out to get some teenagers interested in programming, and thought about a good way to achieve that. A way that allows them to get started with very little friction, build something that’s relevant in their currently life quickly, and avoids frustration.

They were old enough to have their own smartphone, and they were already happily chatting with their friends, using the Telegram messenger. I have already experimented a bit with writing bots for Telegram (e.g. @Umklappbot or @Kaleidogen), and it occurred to me that this might be a good starting point: Chat bot interactions have a very simple data model: message in, response out, all simple text. Much simpler than anything graphical or even web programming. In a way it combines the simplicity of the typical initial programming exercises on the command-line with the impact and relevance of web programming.

But of course “real” bot programming is still too hard – installing a programming environment, setting up a server, deploying, dealing with access tokens, understanding the Telegram Bot API and mapping it to your programming language.

So I built a browser-based Python programming environments for Telegram bots that takes care of all of that. You simply write a single Python function, click the “Deploy” button, and the bot is live. That’s it!

This environment provides a much simpler “API” for the bots: Define a function like the following:

  def private_message(sender, text):
     return "Hello!"

This gets called upon a message, and if it returns a String, that’s the response. That’s it! Not enough to build any kind of Telegram bot, but sufficient for many fun applications.

In fact, my nephew and niece use this to build a simple interactive fiction game, where the player says where they are going (“house”, ”forest”, “lake”) and thus explore the story, and in the end kill the dragon. And my girlfriend created a shopping list bot that we are using “productively”.

If you are curious, you can follow the instructions to create your own bot. There you can also find the source code and instructions for hosting your own instance (on Amazon Web Services).

Help with the project (e.g. improving the sandbox for running untrustworthy python code; making the front-end work better) is of course highly appreciated, too. The frontend is written in PureScript, and the backend in Python, building on Amazon lambda and Amazon DynamoDB.