It was ahead of it's time. Mozilla's A-Frame is kind of it's spiritual successor. VR experiences starting with just a few lines of markup. Check it out https://aframe.io
VRML doesn't need a spiritual successor - it has an actual successor: X3D.
A-Frame looks a bit lightweight to do anything useful. You should have just implemented X3D.
Edit: maybe that's a bit harsh. But I really wonder why you went and created yet another declarative 3D format. X3D is an open standard that already exists, works well in the browser without plugins (X3DOM), and has fairly good support. It's also more advanced than A-Frame, and the syntax is almost as simple.
It's approachable and easy to use. That was the intent. But that doesn't mean its not powerful. It's built on top of a entity-component model (similar in many ways to that of game engines like Unity) that makes it highly extensible and customizable.
The declarative layer is an extremely thin abstraction. A-Frame has feature parity with three.js given its extensibility and entity-component pattern. I'd give it a deeper look.
A-Frame looks a bit lightweight to do anything useful. You should have just implemented X3D.
Edit: maybe that's a bit harsh. But I really wonder why you went and created yet another declarative 3D format. X3D is an open standard that already exists, works well in the browser without plugins (X3DOM), and has fairly good support. It's also more advanced than A-Frame, and the syntax is almost as simple.