The Dion team first got together to make better development tools than the ones we have right now.
But writing tools today is a frustrating endeavour. There are simple problems that can only be solved by giving a heroic effort. For that matter there are important problems that are assumed off the table because even a heroic effort couldn't produce a reliable solution.
We don't think things have to be this hard. Our idea for improving this situation is to reconsider the format for source code.
Text files are the format for source code in the modern software ecosystem. All of our tools are centered around doing something with text files. That can mean editing text files, managing them, sharing them, translating them, generating them, and maybe even more. Every one of these tools is made more useful by the fact that their inputs and outputs can plug into the inputs and outputs of other tools—because they all use text. Without a doubt, text files have worked well as a format for source code.
However, we don't think that text files will be the only format forever. To us it seems that development tools are stuck. In our presentation at Handmade Seattle 2020, we analyzed some aspects of text files that make tool development harder. We believe we have left our source code format underdeveloped for long enough, and that it is one of main things holding back our tools.
When the research and development are done, the Dion Format will be our attempt at upgrading the source code format. We already have a lot of perspectives on what will make a good or bad format, and in our research we hope to gather and share even more insights. If we are successful, an ecosystem will grow around the Dion Format with just as much range as the text ecosystem, and with clearly superior tools.
Right now, this format is our main project. You can see the detailed plan by reading our roadmap, and you can keep up with our progress on our blog.