a version of this existed for a few months last year but it was static. it was HTML with writing and pictures and videos and sounds. i had this feeling that the code should be as important as the content, that structurally each piece in relation to each other piece shouldn't change, that the mazelike quality should emerge from me intricately arranging paths through it. like classic hypertext

this is possible in mazelike research sprints on the internet

so i or you can author smaller fragments that get arranged

I'm trying to picture the scene inside, like I was trying to picture the scene in the tree.