[Simh] Software, firmware, and friends

[Paul Koning]

> On Mar 16, 2017, at 3:02 PM, Seth Morabito <lists+simh at loomcom.com> wrote:
> 
> ... but I
> am seeing results that lead me to believe the manual is incorrect, or
> more likely incomplete. (The manual, after all, is targeted at people
> trying to write software for the WE32100, not people trying to
> implement it, so I can't really blame the authors)

Still a manual should fully describe how the machine operates as seen from the programmer's point of view.

This sort of thing has been around forever and continues.  In recent years I've worked with two different MIPS instruction set SOCs.  We nicknamed the manual for the first of those "The Book of Lies" while the second one was called "The Book of Omissions".  Fun stuff like not describing accurately how to flush one of the caches to memory.

Long ago, an emulator writer in Holland ran into this: Dirk Gruene wrote an emulator for the Electrologica X8, back in 1972.  The idea was to help people debug "bare metal" programs.  Fortunately, he had the real machine to compare with (in fact, the emulator runs on the real machine, as an Algol program).  He describes the oddball undocumented things found at length, fun stuff like address error checking that depends on the address mode used.  Or error cases that don't behave the same way every run.

It's in CWI report MR141/72, unfortunately in Dutch.

	paul