Is TermKit dead? Why?

Such a beautiful demo. I was (am?) really looking forward to this "making it." I do think it's going to take a pretty massive development effort to make it work well though.

For starters, you're going to want to be dealing with objects or structured data rather than raw strings. So you'll be rewriting or writing wrappers for a lot of Unix. Powershell hasn't really been ported to Unix yet, presumably because the "unix greybeards" as he puts it are dogmatically attached to plain text as the communication medium, and not without reason. Maybe if you made HTML or JSON the medium you could simplify this? I suspect you'll want the opacity only live objects can provide though, and that probably means you're opening yourself up to the usual parade of problems.

He mentions that Unix's execution model makes this kind of a non-starter. I have no idea which things here were actually the problem, but I can definitely see how building on very early Node exacerbated the problem. OTOH Node seems like the "correct" platform—Lighttable is also arriving as essentially a browser bundled with embedded Node.

In all, I hope he does pick it up. I would bet that part of why this hasn't happened before or since is because there are many ways to approach the problem that would work but they all probably involve deep changes to the OS—or at least lots of new code. Microsoft can get away with these kinds of things because they also have the whole OS and VM, and they could throw a lot of people at an inelegant solution and still make it work. One guy, working on it in his spare time probably can't produce that volume and there will always be the nagging sensation that if approached from another angle it would be less work. I don't think it's true, but I could see falling into the trap.