Emulation vs. "the real thing" was: Re: Minimal CP-M SBC design

Subject: Re: Emulation vs. "the real thing" was: Re: Minimal CP-M SBC design From: "Chuck Guzis" <cclist at sydex.com> Date: Mon, 19 May 2008 10:00:33 -0700 To: cctalk at classiccmp.org Allison wrote: As you stated, it depends upon what you want to do. 22Nice was written in response to those, who, back in 1986 still had x80 CP/M systems and were looking to migrate to the PC. It was the answer to the question "...But how do I get my old <fill in the application name> to work>?" The goal was not emulation of a Z80/8080 per se, but finding a way to make the transition to the MS-DOS world as seamless as possible. We didn't care about the Osborne/Kaypro/whatever experience or the CP/M experience, only how to get people going on a new platform. By and large, it worked pretty well, with no drivers, TSRs or knowledge about CP/M. You renamed your .COM files, ran GENCOM on them and you were off and running. The SUBMIT capability, for example, was deeemd superflous, as the MS-DOS BAT capability was better in just about every way and simple enough to adapt to. (The first versions of 22Nice ran on a Compupro 85/88 S-100 machine, but that's another story). 22Disk was written solely as a way to get files to 22Nice. We'd considered just making the package read-only, but "in for a penny, in for a pound" thinking produced that bit of creeping featurism. To a CP/M purist, this approach (not providing an isolated emulated Z80 that you could load CP/M into) was probably heresy, but it worked well enough and accomplished the goal. The rise of better PC software eventually relegated 22Nice to the back burner, as that was seen as the ultimate objective.

I'd be lost without uC emulators today--they provide a convenient quick code check without the labor of trying to figure out what the heck is going on in that little block of plastic. But my reason for using an emulator there is very different from that of using 22Nice-- I'm not interested in getting rid of the uC, but getting *to* it.

I'll use an emulator to satisfy my curiosity for hard-to-get old hardware. I've used the SIMH 1620 emulator to scratch an itch in my brain about some code I wrote 40+ years ago, but I would never confuse that with the actual experience of using a 1620 with the blinkenlights, parity checks, clackety console typewriter and glacially slow execution. Nor do I have the will or resources to resurrect or construct my own CADET. Nor do I want one in my office.

Sometimes, I'll use an emulator to figure out how some old piece of software worked, such as WPS-I on an old DECStation. The emulator in this case is better than the real thing, because I can modify the emulator code to show me what's happening internally. This would be at least very difficult on the real hardware--even if I really had it at hand.

This gets back to why I questioned the lack of diskette drives on a "real" Z80 design running CP/M. It seemed to me that if one is after an "experience" and is willing to go to considerable lengths to get it, that it should be as accurate and complete as possible. While 3D computer simulation of skydiving can be made to be very accurate, there's nothing like jumping out of a real aircraft for realism. Having noiseless, crashproof RAM-drives just wouldn't do it for me.

It all depends upon what your objective is.

Cheers, Chuck