Thank you to @mkvlr and @jackrusher and everyone else behind Clerk. I wrote about something I was super stoked to build, something I've wanted to do for 5+ years. It was a magnificent way to help build quickly a recursive descent parser from Markdown, to generate a tree, and get confidence it was correct. I ended up building tests in Koacha (thanks @plexus and team!), to make changes and avoid breaking things, using file watcher, etc. It was a sublime experience. And then slogged away for hours transforming the tree into Vega JSON, again using Clerk, to get fast feedback. (Make change, hit save in Cursive (thank you @cfleming!), watch Clerk update and Kaocha tests run...). I did hit a wall when I finally discovered that Vega wouldn't render on laptop, but rendered in online editor. What the... don't recall how I fixed it.. but the online Vega editor is awesome, too. Automatically renders every time you change the file!) Oh, and thanks to @metasoarous for Oz, which was the gateway drug for using Vega/Vega-Lite with Clojure, and his help 4 years ago. https://twitter.com/realgenekim/status/1520852515998232576?s=21&t=m4mBDGeo-vpkyxjWTn7X2w

...oh, and @djblue 's Portal was instrumental in this, too. How else can you inspect giant tree data structure? (And to think I used to do this using pprint !!! 😭

(I do wish I knew how to make the radial tree not have squiggly lines. Maybe next month. 😁 Radial trees are insane. "Triangular" trees are easy and straightforward. Radial trees actually require specifying a transform (using polar coordinates?) encoded as strings in JSON :face_vomiting:. I suspect some poor graduate student was forced to do that, and was happy to leave the project at the end. 😂