It was something I didn't expect because I started with MS-DOS 2.11 and so it took me awhile to get used to this aspect of CP/M.
I managed to trash a couple disks for the Osborne I bought at a garage sale in the early 90's, which was the first CP/M machine I handled in the flesh. I never really thought of MS-DOS as being "user friendly" but, yeah, coming to CP/M after MS-DOS makes CP/M feel like something a caveman chiseled out of a rock... and then hurled off a cliff at your head.
This is one of the reasons why my personal favorite Z-80 operating system(s) are the various TRS-DOS variations/clones for the Tandy Model I/III; even though there are some aspects of them that are less like MS-DOS than CP/M is (different drive names, etc) they're significantly more forgiving/friendly. Granted they're a lot more technically sophisticated/complex than CP/M. (Like, for instance, how they heavily rely on disk overlays. It's great from a functionality standpoint but it also means the optimal drive configuration for a TRS-80 is *three* disk drives, with a System disk permanently duct-taped into drive 0.)...
You know, in a way I think you could make a case for the PC's success in part the result of it arguably having more in common with a TRS-80 than a CP/M machine. Think about it: Both machines had fairly extensive device APIs and low-level services built into ROM (Level II BASIC and the PC BIOS respectfully), that programs could use to leverage to write significantly richer software than CP/M's limited set of paper TTY-oriented I/O calls allow... yet despite that both ended up with software bases that really rely on "bit-banging"-level compatible hardware because when you're dealing with small low-horsepower computers it's very difficult to achieve with general purpose APIs what you can do with brute force engineering. (And, significantly, when these systems were built nobody really had a good idea what an all-purpose device-independent OS with a "rich" user experience would even look like. I would maybe credit the original Macintosh for being an early low-resource crack at this that had surprisingly good device independence... but also uses way more memory than would have been practical in the 1970's, and was originally devilishly hard to write software for.) "Generic" MS-DOS compatible computers basically completely cratered as a genre after only about three years on the market because of the weaknesses inherent in trying to restrict compatibility to a meager and inadequate API that don't really provide anything but disk access; by 1984 a "PC Compatible" had to be almost as hardware compatible with an original IBM 5150 as a TRS-80 clone had to be with an original TRS-80, or an Apple clone with an original Apple II, etc, to be a sellable product.
MS-DOS may have carried around inside of it some API compatibility with CP/M, but even that
ended up being largely depreciated when the "new" disk API came out with MS-DOS 2.0. So, yeah. I guess I actually don't feel like PCs are all that similar to CP/M machines despite the ancestrial cross-pollination.
Gary Kildall seemed pretty reasonable about settlements too so maybe people would have been persuaded to settle had CP/M hung around, but the biggest insult would have been that it was based on CP/M 2.2, NOT CP/M 3 or MP/M or anything he worked on later which is more of a threat as it represents a fork that would only serve to draw customers away, not back and forth between... MSX-DOS would have been similar. It would have been based on earlier CP/M versions, and hence driven compatability back towards version 1. That in and of itself is a bit of a threat as it lowers the lowest common denominator and invalidates future developments and hence revenue from upgrades.
likewise, if software companies kept pushing out CP/M 1.0 compatible software, no one is going to upgrade their OS.
CP/M 1.0 compatible software? FWIW, 1.4 was the first version that was even remotely common in the wild, and for most practical purposes CP/M 2.2 is universally considered the baseline version. If you compare the
CP/M parts of MS-DOS to the
CP/M BDOS call list you'll see it basically copied the 2.2 API,
as did MSX-DOS. Anyway, I'm kind of curious what you feel like was implemented in CP/M 3 or MP/M would change anything with regard to making CP/M a more "desirable" user environment, specifically with regard to making *good* software more portable?
I think it's worth noting that CP/M really was a dumpster file by the early 80's.
Here's a recollection from someone who worked on the development of the Atari ST in 1984. Atari contracted with DRI for both the GEM GUI and an underlying OS for it to run on top of, and that underlying OS was originally going to be CP/M-68K. Read the article for more details, but even in 1984 the version of CP/M-68K they were working with was stuck with nothing but the rusty old CP/M filesystem calls and character I/O; the team was "saved" from CP/M by switching to a DRI skunkworks project called GEMDOS which, ironically for this thread perhaps, was a
knockoff of MS-DOS, using the FAT filesystem and cloning the better MS-DOS 2.x filesystem API; this is what eventually became the underpinnings of Atari's TOS, not CP/M-68k. (The history of GEMDOS is murky, but it looks like it was related to PC-MODE and other components that turned Concurrent CP/M-86 into Concurrent DOS... and then of course eventually DR-DOS.) So... yeah, it kind of reads like by the mid-80's even DRI was done with CP/M.
For the most-part yes, excepting that there was a slight almost-formed movement of people with z80 systems that *didn't* support CP/M who wanted to use it, because IIRC, Wordstar was still pretty big and hung in there until around 1990 and there were other apps too they wanted, for exactly the reasons I've mentioned. They were tired of losing data and the PC hadn't quite become the new defacto standard until the late 80s.
In other words, owners of machines that had been orphaned by their manufacturers or simply not taken seriously enough by software companies to bother porting "productivity" software to game computers found that hacking their computers to run CP/M at least gave them *something* they could use to stretch their investment a little further. That, in a nutshell, is what CP/M-80 was good for in the latter half of the 1980's. And, sure, Wordstar is definitely better than nothing.
