Today I use an editor that knows, semantically, the context of everything I type. It suggests nouns, and checks compatibility. It automatically shows me documentation for APIs. It knows several hundred common pitfalls to watch out for and suggests corrections for them. It even auto-formats my code so I don't have to dink around doing that.
We've come a long long way in 40 years.
* Ok, it was actually 1979. We got Vaxes and VT100 terminals with VMS EDIT 4 years later. The state of the art in editing was QED, a souped up version of unix 'ed', and vi hadn't really escaped Berkeley yet. Somewhere in MA and CA, some people were munging TECO to become Emacs.
Light Table (http://lighttable.com/) started down the path of what I was thinking, especially with the continual/inline code evaluation.
Once again, these are primarily just thoughts I've been considering. I haven't found a good outlet for discussing my ideas, everyone I discuss them with seems to respond that "that's the way things are" or "that sounds complicated". I've been considering writing a little editor proof of concept to explore some of these thoughts.