This isn’t super “futuristic” but it is work that I did in my spare time that I’m proud of and I think points in a general future direction. I even end the tweet thread with a meta note about the future. I spent my spare time building a visualizer for the new ruby JIT named YJIT. You can check the thread for details, but I did this primarily as a learning experience and it was a great one. I think all code bases deserve a visualization. But today those are hard, hence my meta note. Here I’ll be more direct on the meta point. I think visual programming is awesome. But we can’t leave all the nice things to new languages or wall gardens like smalltalk. We need environments that let us visualize the workings of existing codebases. I don’t think a one size fits all visualization is ever the answer. We need to be able to build visualizations at a higher level of abstraction for our codebases. We need to be able to see the entities of our codebases and how they relate. We need to be able to keep past values of our running programs around. We need an environment that can help us do this regardless of which language we are in. In every size-able codebase I work in I make my own bespoke visualizer. But doing that is tedious and relies on extensive experience building frontends. Does it need to be that way? I think not. https://twitter.com/jimmyhmiller/status/1531638254264541185

Could you reshare your images and gifs here? Twitter downsamples them too much.

Yeah I can do that. Also going to be working on a blog post soonish for this work.

Can only see the rough shapes. Yay, Ruby!

We need to be able to keep past values of our running programs around.

This is huge. Remember when Gary Bernhardt wrote a(n) editor that would use past runs to generate a highlighting mode based on the perf of each line of code? Good times.

Here are the images from the tweets

Looking forward to the blog post with more details on what you’ve built here. 😄