I still want to be able to tell a coworker to "go past the routers, turn to the left and go past views until you get to models; the one you're looking for will

I still want to be able to tell a coworker to "go past the routers, turn to the left and go past views until you get to models; the one you're looking for will be the one on top of the pile on the right". It would be "wow" for sure, but also - spatial memory is one of the easiest to hijack and repurpose for something useful. Of course, being able to visualize the codebase like this would require a Smalltalk-level system (I'm seeing a chance here for Smalltalk & Moose/Roassal/Mondrian/GToolkit/etc.). The problem: once you get to that model, you need to read it. And until now reading lots of text in VR wasn't the best experience, not without things costing $10k+. So, the idea is shelved and waits for a good opportunity in the future... Maybe.