[ajross]

FWIW: really the best example of that kind of thought was OSS's "Action!" language for the Atari computers. Unlike all the modern ideas in this space, this was actually a contemporary product (released in 1983). It's a little like K&R C in terms of semantics, with a bunch of concessions[1] to the limitations of the device.

Alas it came at a time in history where it couldn't make much impact. The Atari platform was being eclipsed by Commodore at the time, and the market for Serious Tool Development in the desktop world was swinging hard toward IBM (Turbo Pascal for MSDOS was arriving right about the same time).

[1] Like a more pascal-ish syntax. Atari didn't have curly braces on its keyboard! Also IIRC "POINTER" types were limited to being statically linked into the zero page, as that matched the operation of the hardware without requiring the runtime to move stuff around for you every time you wanted in do a load.

[zozbot234]

> Atari didn't have curly braces on its keyboard!

Isn't that why C also allows the <% and %> digraphs as alternatives?

[ajross]

Not the Atari itself, which never had a C compiler. But systems like that, yeah. Lots of 70's systems were designed for non-ASCII or ASCII-subset[1] character sets, whereas Unix had made very comfortable use of every funny symbol it could find.

[1] e.g. the original Apple II character ROM had only the 64 characters from 0x20 to 0x5f (no lower case, even) squished into 6 bits of addressing. The keyboard reflected that limitation, as did those from Atari and Commodore. This wasn't rectified by Apple itself until the IIe, though there were 80 column cards on the market from 1980 on that implemented the full