One thing I can tell you: I'll NEVER be satisfied with how my code is organised on a hierarchical filesystem (and don't tell me I can simulate a graph properly

One thing I can tell you: I'll NEVER be satisfied with how my code is organised on a hierarchical filesystem (and don't tell me I can simulate a graph properly with symlinks). I think some sort of tagging filesystem would be appropriate. Also, I hate how most programming tools assume I want to choose a "canonical" representation of code into files. I'd want to mostly don't care how the code is organized on disk and then write queries which let me view and edit code coming from multiple different files at once. As a completely hypothetical, contrived example, let's say I was still doing Java and had a big class with many getters and setters, I don't want to choose whether I want the getter and setter of each attribute organised in pairs or whether I want all the getters first and then the setters. I want to be able to write a query that will constitute a new view, so if I want either option I just have to switch views. A third view would simply be the physical representation on disk, which is still needed from and dependencies-loading standpoint. This way of working would truly shine in much more complex scenarios but I hope you see what I mean.