"CP/M compatible" vs. "MS-DOS Compatible" machines?

There were also the independent implementations such as TPM (for the QX-10) and TuboDOS. There were others, even a a couple that provided the capability to run CP/M 2.2 programs under non-CP/M OS Z80 platforms. CP/M, particularly 2.2 was easily cloned. Most applications used the information in the reference manual set and didn't rely on "undocumented" features. IIRC, the bigger problem when MP/M came along, was trying to work with those programs that were just plain sloppy and performed file open requests without closing the file (since CP/M didn't even keep so much as an open file count, you could be very sloppy, even opening a file multiple times or not at all (as long as you filled in the FCB fields correctly). Some programs saved FCBs and then used them later. I wanted to add a "put that diskette back" alarm to my CP/M implementation when there were files opened for writing (one needn't even turn on the drive motor--just periodically sample the write- protect status; a diskette being inserted or removed will toggle it a couple of times). I'd already done it for DX-85M and it reduced our diskette file corruption problems to nearly zero. That "R/O error" diagnostic from CP/M was unreliable at best and nearly useless even when it worked. It was then that I discovered that CP/M made no attempt to track open files and indeed, the behavior of many programs depended on this. Cheers, Chuck