MIT want to make $12 Apple II clone

Chuck> "Incremental compilation" to pseudo-code has been around for a Chuck> long time--at least from the 1960's. One advantage is that Chuck> your pseudo- code can have features that the native machine Chuck> architecture lacks, such as virtual memory, relocatability, Chuck> large memory space, etc. Chuck> If done well, the speed can rival (and in some cases, exceed) Chuck> that of direct-to-native code. I know of a fellow who did a Chuck> COBOL compiler and run-time back in the 60s that worked inside Chuck> of a CDC 6000 series PPU (4K of 12 bit words). (Was there ever Chuck> a COBOL for the 4K PDP 8/L?) Are you sure that COBOL ran on a PPU? I can't imagine a reason for doing that, other than as a "watch this" stunt. On the other hand, Cobol on a 6000 CPU would be sensible and useful. Chuck> ...A few years later, I used the same technique to write a Chuck> business BASIC that compiled to an intermediate stack-machine Chuck> code that ran on an 8085--and multitasked with no special Chuck> hardware. Was that before Forth came out? That's the classic example of a p-code language with multitasking. Another earlier example is BASIC-PLUS for the PDP-11. That's an incremental compiler to P-code. Also, in the earlier versions (RSTS-11, for the no-MMU PDP-11s) it included multitasking (timesharing) support in the P-code machinery. The later versions (RSTS/E) didn't need that because there was a real kernel with MMU hardware. paul