One of the first systems that I could actually touch was a 68K/S100 system back in early '80s; it ran a unix-like OS. It was made by a Seattle area company named Empirical Research Group. The CPU board had Forth in ROM. I was lucky enough to witness one of the designers perform some serious diagnostics on other boards in the system using only the CPU/Forth.

I don't think they were the first to come up with this idea.

On Wed, Jun 20, 2018 at 10:39 PM Kurt H Maier <khm@sciops.net> wrote:

On Wed, Jun 20, 2018 at 10:35:42PM +0100, Ethan A. Gardener wrote:
>
> a sort of operating system where the primary interface to all tasks is
> a Forth interpreter.

I think we've talked about this in another venue some years back, but I
often thing of the OpenFirmware implementation used by the OLPC XO-1
laptop. Instead of a BIOS or UEFI or linux trash in their stead, the
system was managed by an OpenFirmware installation, much of which was
written in Forth, and whose primary interface was a Forth shell. This
environment had complete access to the hardware of the system, which
was used by the project to create really comprehensive hardware
diagnostics tools.

I mostly used it for screwing around, but it was fairly complete; it
supported the wifi hardware and the webcam, and I often thought I'd like
a computer that just booted into this environment and stayed there. I'm
glad to hear you're still experimenting along these lines. There's a
lot of value in a system whose primary interface is the programming
environment. I work with computers because of the Commodore VIC-20...
and I wonder if I'd have ever given a damn about the field if my first
exposure to computers involved a Modern User Experience.

khM