dyer (07/14/82)
Groan... My point, condensed for the telegraphic mode of netnews, was: It became popular to port UNIX to new machines after the people at Bell showed that the whole slew of UNIX C code would still run on a Interdata, once the kernel was moved. This is because most application code was written in C, and performed non-trivial interactions with the kernel through system calls (i.e. just porting a C compiler to another system doesn't usually give you the virtual OS support that most programs assume.) There's so much good stuff in the UNIX package, that the effort involved to port it to another machine makes it all worth-while. (Not to mention how nice UNIX is even without all the extra software!) NOW, most CP/M software can be split into three categories: 8080/Z80 assembler code (largest group) CBASIC or Microsoft BASIC programs (a bit smaller) all the rest (C, PASCAL, whatever, it's not a lot) All the assembler stuff has to be rewritten. Period. The BASIC stuff mostly never cared that it ran under CP/M. PLUS that BASIC interpreter (written in assembler!) has to be rewritten. Some high-level stuff interacted with CP/M thru "bios()" calls, but it's such a small amount that it's hardly worthwhile. My point is that you don't gain anywhere near the same treasure- trove of software when you port CP/M to another machine architecture. All you really give the user is an identical user interface, and one that isn't so hot. If you really wanted, you could put a CP/M like shell on top of a more modern 68000 operating system, and you'd have gained 95% of all the advantages of a complete port.