• Please review our updated Terms and Rules here

How long did CP/M have a hold before MS-DOS became a thing

A

alank2

Veteran Member
Joined
Aug 3, 2016
Messages
2,264
Location
USA
I've had it in my mind that each had its own "era" in a sequential thing like "CP/M was it for 8 years, and then MS-DOS was it for X years", but the early date of the IBM PC in 1981 seems to throw that idea out the window. Obviously CP/M was earlier, but how much earlier and for how long did it "rule"? What makes me question this is both that (1) What CP/M machines propelled it to popularity before the release of the PC specifically. What I've noticed is that there seems to be a lot of PC's that were CP/M based, but AFTER that 1981 year. Was it a situation where the PC hadn't taken over and some of CP/M's momentum from before 1981 drove the development of those systems even after the PC was released?
 
Applications for CP/M were a major factor which could be dated starting with Wordstar in 1978. It declined slowly. The DEC Rainbow was designed to run CP/M-80 programs since no one expected applications to be moved to MS-DOS as fast as happened*. Mid-80s would be when CP/M fell behind and mostly remained in use by those who already had an install or purchased one of the small number of high end budget Z-80 machines introduced later.

* I remember one of the magazines including a simple full screen text editor because of the lack of any for MS-DOS. By the time the article was published, there were dozens of commercial text editors available.
 
One of the things that I've learned when doing a "deep dive" into any historical subject is that the concept of "era" only means "when X was dominant". There were many years where the old was still in wide use while the new was growing.

So there's going to be the "era of CP/M" where CP/M was dominant. When the IBM-PC was being developed, CP/M was going to be used, but IBM ran into issues, and wanting to get their system out the door fast, chose MS-DOS. But CP/M was still very dominant well into the MS-DOS era.

I think the problem that we are having is not really knowing about computers pre-PC era. The Altair 8800 didn't come out until 1975 (and probably couldn't run CP/M until a year or so later). So CP/M must have been established before then. But probably not in a market that we would recognize as the "personal computer market".
 
rlauzon said:
One of the things that I've learned when doing a "deep dive" into any historical subject is that the concept of "era" only means "when X was dominant". There were many years where the old was still in wide use while the new was growing.

So there's going to be the "era of CP/M" where CP/M was dominant. When the IBM-PC was being developed, CP/M was going to be used, but IBM ran into issues, and wanting to get their system out the door fast, chose MS-DOS. But CP/M was still very dominant well into the MS-DOS era.

I think the problem that we are having is not really knowing about computers pre-PC era. The Altair 8800 didn't come out until 1975 (and probably couldn't run CP/M until a year or so later). So CP/M must have been established before then. But probably not in a market that we would recognize as the "personal computer market".
Click to expand...
Yes, overlapping curves, each technological "era" increases in market share as the priors decline. In the case of DOS vs CP/M, it was obvious by DOS 2.0 1983 with directory trees, DOS was dominating as DRI tried to play catch up. Also, IBM practically giving away $40 DOS with more utilities+features beats paying $240 CP/M-86 with fewer utilities. I am contrasting the early 1980's marketing strategy of DOS vs CP/M; I am a big DRI CP/M fan and it has design elegance, efficiency, and shines among the 8bit computers, even today.
 
Last edited:
ja1001200 said:
Yes, overlapping curves, each technological "era" increasing in market share as the priors decline. In the case of DOS vs CP/M, it is obvious by DOS 2.0 1983 with directory trees, DOS was dominate as DRI tried to play catch up. Also, practically giving away $40 DOS with more utilities+features beats paying $240 CP/M-86 with fewer utilities. I am contrasting the early 1980's marketing strategy of DOS vs CP/M; I am a big DRI CP/M fan and it has design elegance, efficiency, and shined among the 8bit computers.
Click to expand...

Yup. Directories and price. Those are probably the two things that caused CP/M to go down.

When the IBM-PC came out, the low end systems were supposed to be "boot to BASIC and use a cassette", but that era was already ending. Heck, even my high school had floppy drives for their Commodore PETs.
So DOS 1.0 and CP/M were pretty much even.

But when disk capacities increased, and hard drives became more affordable, CP/M really started to show its cracks. CP/M's user spaces always felt like a kludge to me and certainly were harder to work with than directories.

And the licensing price was a bad idea. Gary didn't realize that the market has shifted from companies buying these to home users buying them - and $200 meant a lot.

[Side note: Microsoft made the same mistake when the Asus eee PC came out and wanted to charge $100 for a Windows license while Linux was free.]
 
If MS cut Asus eee PC a deal with Windows license, then they would need to cut a deal with everyone else and lose $$$$.

The only time I recall MS giving out cheap licenses was for XP in China (which pirated 99% of OS's anyway) and for rebuilders selling refurb pcs which MS got paid for already anyway.

I don't think home users cared about the difference in price between CP/M and DOS because they could just pirate it anyway. The thing is people tended to have a computer at home (well anything not from Atari or Commodore) to do work so they just purchased what they used at work and DOS machine just dominated there (OEM licenses were much cheaper than retail).
 
As I remember, all of the Tandy's, even my handheld at least had BASIC and everything else on up the line came with LDOS and maybe a few apps.
 
rlauzon said:
But when disk capacities increased, and hard drives became more affordable, CP/M really started to show its cracks. CP/M's user spaces always felt like a kludge to me and certainly were harder to work with than directories.
Click to expand...
Yeah, when it came to HD, CP/M paled in comparison.

DRI CP/M 2.2
USER non-hierarchical single level, max 16 directories​
files may have upto 65,535 records x 128 bytes each​
filesize up to 8M​
total files 2048​
harddisk upto 8M​
total drives or logical partitions upto 16​

DRI CP/M Plus
USER non-hierarchical single level, max 16 directories​
files may have upto 65,535 records x 512 bytes each​
filesize upto 32M​
total files 2048​
harddisk upto 512M​
total drives or logical partitions upto 16​

However, most applications only handle filesize up to 512K due to DRI not documenting the S2 byte (chunk blocks) that count in multiple of 512K chunks.

Z-System ZxDOS for CP/M 2.2 allows filesize up to 32M, harddisk up to 32M and up to 65,535 records x 512 bytes.

These are my notes and I know my limits, so I am willing to learn if you someone shares better insight.
Jim
 
Last edited:
Had Gary Kildall lowered his O/S price to undercut Microsoft when the IBM PC launched it would have stuck around a lot longer then it did.
 
And CP/M probably would have developed all the features that later variants of MSDOS did in that case.

Dave
 
CP/M was dominant pretty much for a decade, from 1974 to 1984. By 1985, DOS was picking up. CP/M hung around until the end of the decade, but no new computers were coming out.

It was killed primarily by two things. The 286 and the x87 coprocessor. PCs were backwards compatible and compatible with each other. CP/M not so much, but by '86, clone 286 machines were common and really performed well. By 87, the future was clear to everyone, and 386 was around the corner.

Few CP/M machines were built after 1984, though some stuck around until the end of the decade, when they too could no longer compete either on price or power. Amstrads probably stuck around the longest due to being low priced word processors.
 
Plus some installed CP/M machines stayed around because of the cost of changing (i.e. process, time, effort, monetary costs) - the effect of the installed base.

A place I worked at around '88 had just got recent 286 based diskless PCs, connected to Netware, using the monolithic NET3 client. Those machine probably were 640k, and EGA (if I recall correctly). This was a manufacturer of turnkey electronic devices. Those devices had previously been Z80 based, but the current batch when I joined were x86 (NEC V20) based. Hence having Z80 based development machines was useful if only for easy testing w/o having to use the target devices.

The development department had changed to PCs maybe a year before I joined, the manufacturing and test department still used some CP/M machines (Morrow MD 10/20?). They stopped using those last machines while I was there, and I got a couple of the old ones, which I eventually scavenged for parts for building my own 286 PC. I also acquired a whole bunch of disks with development tools.

Another such turnkey device company where I worked from '90 only used PCs. A lot of big AT clones, but with 512 or 640k RAM, and a few new 386sx machines with 1M RAM, all again using Netware. They did however still run some 8080/Z80 CP/M software using simulators on the PCs. Those few programs were replaced while I was there.

So in that field, I'd suggest CP/M use died out between '87 and '88, with only some s/w lasting until '90.
 
Agent Orange said:
As I remember, all of the Tandy's, even my handheld at least had BASIC and everything else on up the line came with LDOS and maybe a few apps.
Click to expand...
The TRS-80 line didn't like CP/M very well due to the memory layout (CP/M expected things to be in certain places and the TRS-80's didn't do that). I think we had to wait until the Model 4 came out to get CP/M. I know that I have a version that runs on my 4P.
 
furball1985 said:
Had Gary Kildall lowered his O/S price to undercut Microsoft when the IBM PC launched it would have stuck around a lot longer then it did.
Click to expand...
But part of the problem is that Gary didn't have CP/M-86 when IBM needed it.

The second part of the problem is that CP/M-86 is not compatible with CP/M-80, so owners would have to re-purchase all their software.

So the choices were
1. Stick with the Z80 and CP/M-80, which was harder to do once the 286 came out
2. Upgrade to CP/M-86 or MS-DOS and re-purchase your applications. And applications for CP/M-86 were pretty scarse, but apps for MS-DOS were easily purchasable.
 
rlauzon said:
The TRS-80 line didn't like CP/M very well due to the memory layout (CP/M expected things to be in certain places and the TRS-80's didn't do that). I think we had to wait until the Model 4 came out to get CP/M. I know that I have a version that runs on my 4P.
Click to expand...
Pickles and Trout had a version of CP/M for the Model II line starting in 1979.
 
JimA1 said:
Interestingly enough, I was just discussing "what if Tandy supported CP/M sooner", even as late as 1980. Imagine the potential of the TRS80 with their largest marketshare, combined with DRI CP/M during its highest dominance. Sounds like a hind-sight win-win missed opprutinity.

In case someone interested, please see wishing the Model 4 was introduced in 1980 instead of the Model III.
Click to expand...

CP/M already had a huge share. Tandy would have extended it, and maybe cemented a few more years, but as we discussed in another thread, the rate at which x86 sped up while being backwards compatible, the fact that IBM came into the picture, that Microsoft took over DOS and that the 8087 existed was pretty much the end of z80.

The lack of graphics for cp/m also killed it as did zilog's poor decisions at the time.

It's not unreasinable to assume that had DRI been chosen instead of Microsoft, we'd all be using either MACs or Commodores today.
 
cj7hawk said:
The lack of graphics for cp/m also killed it as did zilog's poor decisions at the time.
Click to expand...
Most CP/M systems were designed similar to the way minicomputers were designed: A computer part, and a terminal part. The fact that they were often in the same box made people forget that.
So what most of these systems lacked was a way to access screen memory since the screen memory was **not** part of the computer and, simply, not directly addressable from the computer.

The later Kaypros, though, did have some graphics capabilities where they could set blocks on/off (similar to the TRS-80's SET/RESET) as well as create lines. But they were done by sending commands to the "terminal" part of the computer.

cj7hawk said:
It's not unreasonable to assume that had DRI been chosen instead of Microsoft, we'd all be using either MACs or Commodores today.
Click to expand...
These kinds of what-ifs are always interesting.

In my take, if DRI had been chosen for the IBM-PC, and DRI had started to license it for a reasonable price, the clones would have taken off pretty much the same as they did under MS-DOS.
That would have pushed the computer/terminal model forward instead of MS-DOS's "I control all the hardware" model.
That would have led us to MP/M sooner. So the potential of having a multi-user, multi-tasking OS early on is also not unreasonable.
For graphics, we might have had something similar to X-Windows, but in hardware since that would have gone with the computer/terminal model that was "standard".
 
Japan stuck with CP/M for longer than the USA. The Epson QX-10 came out in 1983 and it had a Z80 and ran CP/M 2.2. A little later they released the QX-16 a dual CPU (z80 & 8088) that ran either CP/M or MS DOS 2.2. So by then the handwriting was on the wall. Of note, the QX-16 does not use PC format disks since it was released just before the IBM PC came out. While it runs MS DOS, it is not IBM PC compatible for IBM only programs or graphics. So it too died a rapid death.
 
You must log in or register to reply here.
Back
Top