My current plan is to generate this kind of graph from documentation pages, similar to what I did with Wikipedia. The dynamic I thought was loading new pages as I open nodes (the same way you can visit the whole Wikipedia by clicking links, you'd be able to open all Wikipedia graph by expanding nodes). I also thought about showing the evolution of the graph over time, like gource.io. But if you are thinking about a real time update that doesn't require a page refresh, it's not in my plans no.
HN user
smokeonline
It's a very good example of a front-end REPL. I used to think hot reloading was good, but to keep the state between reloads is a whole new level.
This idea passes through BDD, Ubiquitous language, code as documentation and single source of truth.
Do you think is possible and desirable to have code so idiomatic that a domain specialist would be able to understand a functionality by looking at the code?
This is a nice idea, I've been looking for things like this to read. But my interest is more about generic systems, not that much about end-user software.
You are talking about "trade-offs" and I had a similar idea about "Values". With this values you would be able to derive the "Rules" (the HOW you mentioned) depending on the context and chosen trade-offs.
About it I think the 17 unix philosophy rules are very interesting. https://en.wikipedia.org/wiki/Unix_philosophy#Eric_Raymond%E... and I also started to write a blog about it https://github.com/JpOnline/Blog/blob/master/language.md