sorting 3921

Well: In FP Brookss' introduction to automatic data processing book, he discusses the IBM mod 30 internal micro instructions and has some exercises on that. This could be one language. It would also satisfy another goals of discussing Unified external appearance architecture realized by means of microprogrammed implementation. Thus you cover the distinction between architecture, implementation, realization (i.e. electronics used). The exercises are for example to program some of the 360 instructions in the microcode, and are done as paper and pencil exercises without (needing) an emulator. As a second language you could choose one for a three address field machine. e.g. Ferranti Ferrut, with a magnetic drum (and mercury delay line) memory. The machine language had the form OPcode, Op1, Op2, Next-instr . The programs were buttembled using SOAP the Symbolic Optimum buttembly Program. The third might really be a pair of different languages for a single architecture. I think that a one address and a two address buttembly language is available for one of the Intel 8080, 8085 or perhaps its the Z80. After that the field gets wider. Perhaps PDP-11 with variable length opcode and the problems that causes for instruction decode, vs 360 (original) where the first two bits IL give instruction type, the opcode is 1 byte and otherwise the ISP is quite regular. One can then take an aside to show how this has been elaborated lately. Or VLIW type of machine code vs. RISC machines and their code with skip instructions to handle delay situations. And there's always Knuths' MIX. (I did not check the details for the above, so I apologize for mistakes in details: names etc. Please correct me. It will make for interesting postings) -- Rostyk