Thanks to mike who commented and pointed me to the domain/object project.

The similarities to what I built 20 years ago when I was a teenager are amazing.

I was more concerned with the mechanics of how the data-flow was implemented and given the computing constraints of the time this was an overriding factor. As far as I can tell my final solution is still unique.

Concurrency was the primary goal with an aim to developing a distributed operating system. MPOS (Massively Parallel Operating System). The system was usable enough for me to cluster all the computers at the office where I worked and develop a functional ray-tracer, a back-prop neural net trainer and some fun experiments with genetic algorithms.

Funnily enough 5 years ago I got some funding and wanted to take the idea commercial and create a full implementation as part of a scripting language, but it was deemed to be 'crazy' :)