Why did DOS/86 overtake CP/M-z80 ?

[Eudimorphodon]

You mentioned some longer timeframes for the C64 and I was wondering if you knew what they were as a percentage so we can understand the significance of those numbers with relation to the total number of games released for the C64. If you don't know the figures, a guess is fine. It will still give me more context into what you're saying.

Why are you asking me this? If you want to support your argument you can do the research yourself; MobyGames is a public database. But, sure, I'm still on break so I'll take a quick crack at it, why not.

I guess I get the case you're trying to construct here, but it falls flat on its face, sorry. Let's stick with using the Commodore 64 as the comparison; it seems fair enough, given they were both introduced at nearly the same time (late 1981 for the PC-88, vs. mid-1982 for the C64) and the hardware was produced into the 1990's. Here's an overlaid graph of the number of games produced for the systems between 1982 and 1993; Mobygames only lists a *single* title of the PC-88 from 1981 so we're not really short-changing it by leaving 1981 out.



Let's also chuck into this discussion the *total* number of games MobyGames lists for each platform:

PC-88: 556
C64: 5,598

(It's kind of spooky how close to exactly 10 times the multiple between C64 and PC-88 games is. Wow...)

The thing you seem to be harping on is the question of how many of the platform's total games were produced after 1990? Well, that's easy for the PC-88; if I leave out the *1* game produced in 2023 (which seems reasonable in this context) we have 57+20+3=80/556 = 14.4%. The C64 is tougher because its game development didn't *really* dry up until the late 1990's; to make it as fair as possible, though, I'm arbitrarily cutting it off at 1993, the year "Lemmings" was ported to it *and* the last year that the number of games per year was over 100. (This is pretty arbitrary, though, given 1994 had 87 games and 1995 had 71. If I set the threshold as proportionally low as the "3" I'm allowing for PC-88, IE, only 3% of the output of the peak year, we'd get as far as 1996 with 32 games.) 319+280+107=706/5589 = 12.6%. (If I apply the same cutoff I'm allowing for the PC-88 then we get 896/5,598 = 16%)

These numbers show your premise regarding the PC-88 as being some kind of late-surviving miracle machine is utterly false. PC-88 did not get significantly more development after 1990 than a machine very popular in the West, and in fact died both earlier and far more definitively than the C64. Like I said, the only whimper of an argument you can form when you visualize the data like this is the Commodore 64 started declining significantly from it's "peak" a few years earlier, but the flip side of this is that the C64's numbers are both much larger than and a lot more "spikey" than the PC-88's.

EDITOR'S NOTE: After running these numbers I stumbled across this page, which lists a total of 838 games for the PC-88 and has a "releases per year" chart. I'm a little leery mixing data from different sources, but, sure, If I redo my graph with these numbers it looks like this:



The amusing thing here is counting more games actually makes your case weaker... by a lot. Look how similar the profiles of the curves are now, and if we redo the math for "percentage of games released after 1990" we get 48+20+4=72/838 = 8.6%. The PC-88 is officially now worse than the C64 in terms of late relevance by a mile; the C64 had half again more games released for it after 1990, percentagewise, than the PC-88.

EDIT: I just realized this other source also has platform information for the C64, which likewise lists more games than MobyGames, but I *really* can't be bothered to update the graph again. I did eyeball the corresponding releases per year stats, however, and if you add up the totals for 1990-1993 and divide them by its total of 6316 total games the C64 scores 15.6% of its total releases after 1990, which is even better than MobyGames' numbers; the C64 had, statistically, twice as many games released in this late period than the PC-88.

Though if your answer is "because we never got porn on CP/M systems in the US" well, it's a different perspective - I'll grant you that -

You might think you made a joke here, but, no, this is *exactly* why nobody cared about CP/M. Unless you're running a dental office or accounting firm there was absolutely nothing interesting about it.

In focusing on the PC-88 you couldn't better demonstrate how much you're missing the forest through the trees: The PC-88 "put up a fight" not because it was a CP/M machine, but because it was a GAME MACHINE. First and foremost that's what it was, and I'm kind of baffled that you keep missing how unimportant CP/M was to its success. The very articles you've cited in trying to point out how CP/M "lasted longer" in Japan are very clear about how the system was kept on life support forever because it was a popular game machine, not because it could run archaic mailing list management software from the 1970's. CP/M wasn't even the primary OS of the machine; it had its own DOS that integrated with its MS BASIC, and a lot of these games you're obsessing about were self-booters.(*)

(* If we want to be nitpicky I'd argue that these Japanese computers like the PC-88 and even MSX are more like TRS-80s than they are typical CP/M machines, in that in order to be usable without disk drives their primary OS is Microsoft BASIC that they're happy to boot into from ROM. The fact that their memory maps can accommodate CP/M is almost an accidental consequence of the fact that they incorporated memory paging from the outset, largely for other reasons involving language and graphics support.)

A gaming machine needs CP/M like a fish needs a bicycle. The fact that some of these obscure Japanese gaming machines happened to have a bicycle in the garage is because they were descended from computers where it seemed like a good idea at the time, not because it made even a lick of sense for the niche they ended up falling into. The games that were the only reason people cared about these things as long as they did were not "CP/M games" and certainly weren't transferrable between systems. (You think it's hard to patch Wordstar for a different terminal, just wait until you try to run your PC-88 porno adventure on your Kaypro.) The C64 was more successful than the PC-88 despite Commodore having to, *gasp*, write their own DOS, because (in addition to being a lot cheaper) nobody in the market for a game computer gave a rat's rear about what operating system was used behind the scenes to shovel digital boobs from the floppy disk into their eye holes.

I mean, let's get real about CP/M: there's not a lot to it. Apple DOS is roughly as comprehensive as CP/M and TRS-DOS is significantly fancier than it, and both were pretty much written by one guy. (Paul Laughton and Randy Cook, respectively.) Once a computer company is selling more than a few thousand units there's a lot be said about just hiring someone in-house to maintain your OS instead of having to send royalty checks to DRI. CP/M made sense in a world where computer consultants were hired to assemble practically bespoke business computers out of a menu of individual parts while maintaining some limited pretense of interoperability and the target functionality still might involve paper teletypes, but the case for it did little but erode after that.

 

[danielbooneamerica]

guys, lets just all be happy we finally got the PC and online porn.

Still trying to get porn on TRS-80 to work.

 

[danielbooneamerica]

guys, lets just all be happy we finally got the PC and online porn.

Still trying to get porn on TRS-80 to work.

I am so sorry, commodore had pics also....I now remember someone showing me a nude pic on a C64 about 1988!

 

[Eudimorphodon]

Still trying to get porn on TRS-80 to work.

Yeah, that was a definite drawback to the TRS-80; if you wanted to digitally get your rocks off in front of a CRT you were pretty much forced to invest in an Atari 2600 and a copy of Custer's Revenge.

I am so sorry, commodore had pics also....I now remember someone showing me a nude pic on a C64 about 1988!

Don't forget, there were *tons* of naughty ASCII art pictures floating around on BBSes going back to, well, pretty much since modems were a thing. And, well, there's a reason why Compuserve invented the GIF format. Of course to really make use of those you needed a printer... and a deep, sad dose of desperation, I guess.

 

[danielbooneamerica]

Yeah, that was a definite drawback to the TRS-80; if you wanted to digitally get your rocks off in front of a CRT you were pretty much forced to invest in an Atari 2600 and a copy of Custer's Revenge.



Don't forget, there were *tons* of naughty ASCII art pictures floating around on BBSes going back to, well, pretty much since modems were a thing. And, well, there's a reason why Compuserve invented the GIF format...

i worked on a DEC PDP 11/45 in 1983 that had a huge collection of character based porn.

It was inredible how detailed and real it could look just made up by characters...someone spent a lot of time.

There was a Santa with sled and raindeer and when printed on line printer it must have been 6-8 feet long.

One big regret I have is I did not preserve copies of these files. It is something you had to see and very detailed.

All made with X O . ; - ) * etc

 

[danielbooneamerica]

Why are you asking me this? If you want to support your argument you can do the research yourself; MobyGames is a public database. But, sure, I'm still on break so I'll take a quick crack at it, why not.

I guess I get the case you're trying to construct here, but it falls flat on its face, sorry. Let's stick with using the Commodore 64 as the comparison; it seems fair enough, given they were both introduced at nearly the same time (late 1981 for the PC-88, vs. mid-1982 for the C64) and the hardware was produced into the 1990's. Here's an overlaid graph of the number of games produced for the systems between 1982 and 1993; Mobygames only lists a *single* title of the PC-88 from 1981 so we're not really short-changing it by leaving 1981 out.

View attachment 1270774

Let's also chuck into this discussion the *total* number of games MobyGames lists for each platform:

PC-88: 556
C64: 5,598

(It's kind of spooky how close to exactly 10 times the multiple between C64 and PC-88 games is. Wow...)

The thing you seem to be harping on is the question of how many of the platform's total games were produced after 1990? Well, that's easy for the PC-88; if I leave out the *1* game produced in 2023 (which seems reasonable in this context) we have 57+20+3=80/556 = 14.4%. The C64 is tougher because its game development didn't *really* dry up until the late 1990's; to make it as fair as possible, though, I'm arbitrarily cutting it off at 1993, the year "Lemmings" was ported to it *and* the last year that the number of games per year was over 100. (This is pretty arbitrary, though, given 1994 had 87 games and 1995 had 71. If I set the threshold as proportionally low as the "3" I'm allowing for PC-88, IE, only 3% of the output of the peak year, we'd get as far as 1996 with 32 games.) 319+280+107=706/5589 = 12.6%. (If I apply the same cutoff I'm allowing for the PC-88 then we get 896/5,598 = 16%)

These numbers show your premise regarding the PC-88 as being some kind of late-surviving miracle machine is utterly false. PC-88 did not get significantly more development after 1990 than a machine very popular in the West, and in fact died both earlier and far more definitively than the C64. Like I said, the only whimper of an argument you can form when you visualize the data like this is the Commodore 64 started declining significantly from it's "peak" a few years earlier, but the flip side of this is that the C64's numbers are both much larger than and a lot more "spikey" than the PC-88's.

EDITOR'S NOTE: After running these numbers I stumbled across this page, which lists a total of 838 games for the PC-88 and has a "releases per year" chart. I'm a little leery mixing data from different sources, but, sure, If I redo my graph with these numbers it looks like this:


View attachment 1270775
The amusing thing here is counting more games actually makes your case weaker... by a lot. Look how similar the profiles of the curves are now, and if we redo the math for "percentage of games released after 1990" we get 48+20+4=72/838 = 8.6%. The PC-88 is officially now worse than the C64 in terms of late relevance by a mile; the C64 had half again more games released for it after 1990, percentagewise, than the PC-88.

EDIT: I just realized this other source also has platform information for the C64, which likewise lists more games than MobyGames, but I *really* can't be bothered to update the graph again. I did eyeball the corresponding releases per year stats, however, and if you add up the totals for 1990-1993 and divide them by its total of 6316 total games the C64 scores 15.6% of its total releases after 1990, which is even better than MobyGames' numbers; the C64 had, statistically, twice as many games released in this late period than the PC-88.




You might think you made a joke here, but, no, this is *exactly* why nobody cared about CP/M. Unless you're running a dental office or accounting firm there was absolutely nothing interesting about it.

In focusing on the PC-88 you couldn't better demonstrate how much you're missing the forest through the trees: The PC-88 "put up a fight" not because it was a CP/M machine, but because it was a GAME MACHINE. First and foremost that's what it was, and I'm kind of baffled that you keep missing how unimportant CP/M was to its success. The very articles you've cited in trying to point out how CP/M "lasted longer" in Japan are very clear about how the system was kept on life support forever because it was a popular game machine, not because it could run archaic mailing list management software from the 1970's. CP/M wasn't even the primary OS of the machine; it had its own DOS that integrated with its MS BASIC, and a lot of these games you're obsessing about were self-booters.(*)

(* If we want to be nitpicky I'd argue that these Japanese computers like the PC-88 and even MSX are more like TRS-80s than they are typical CP/M machines, in that in order to be usable without disk drives their primary OS is Microsoft BASIC that they're happy to boot into from ROM. The fact that their memory maps can accommodate CP/M is almost an accidental consequence of the fact that they incorporated memory paging from the outset, largely for other reasons involving language and graphics support.)

A gaming machine needs CP/M like a fish needs a bicycle. The fact that some of these obscure Japanese gaming machines happened to have a bicycle in the garage is because they were descended from computers where it seemed like a good idea at the time, not because it made even a lick of sense for the niche they ended up falling into. The games that were the only reason people cared about these things as long as they did were not "CP/M games" and certainly weren't transferrable between systems. (You think it's hard to patch Wordstar for a different terminal, just wait until you try to run your PC-88 porno adventure on your Kaypro.) The C64 was more successful than the PC-88 despite Commodore having to, *gasp*, write their own DOS, because (in addition to being a lot cheaper) nobody in the market for a game computer gave a rat's rear about what operating system was used behind the scenes to shovel digital boobs from the floppy disk into their eye holes.

I mean, let's get real about CP/M: there's not a lot to it. Apple DOS is roughly as comprehensive as CP/M and TRS-DOS is significantly fancier than it, and both were pretty much written by one guy. (Paul Laughton and Randy Cook, respectively.) Once a computer company is selling more than a few thousand units there's a lot be said about just hiring someone in-house to maintain your OS instead of having to send royalty checks to DRI. CP/M made sense in a world where computer consultants were hired to assemble practically bespoke business computers out of a menu of individual parts while maintaining some limited pretense of interoperability and the target functionality still might involve paper teletypes, but the case for it did little but erode after that.

Hey when you are not busy could you make some of these charts for TRS-80?

 

[cj7hawk]

Why are you asking me this? If you want to support your argument you can do the research yourself; MobyGames is a public database. But, sure, I'm still on break so I'll take a quick crack at it, why not.

I guess I get the case you're trying to construct here, but it falls flat on its face, sorry.

I think you're forgetting the question I posed at the start....

But it also seems like people had a preconceived idea by the early 80s of what a CP/M machine or z80 based system had to be - it's not like there was even a fight. The z80 just threw in the towel, long before other matters such as compatibility were key differentiators. Perhaps the only people who wanted S100/z80/CP/M by then were Spectrum owners, with an architecture that couldn't get much further away from that ideal if it tried.

That's a question. Not a construction. Everything after it is me repeating back what is learned from this thread. Your input is appreciated. I'm enjoying the information I'm learning from this thread.

I'm not suggesting these machines survived because they were CP/M machines... I'm wondering why these machines survived in Japan, Eastern Europe, Western Europe etc, but died out almost like they were hit by a meteorite in the US, then it spread rapidly. By 1990, the performance of x86 was enough, but early-mid 80s it's not like z80 based CP/M systems even tried to fight back. Companies either went directly PC or died out entirely. With the exception of Apple, who stuck to their architectures.

View attachment 1270774

That's a pretty interesting graph. Makes me wonder what happened in 1987 to cause a resurgance for these machine that CP/M never took advantage of.

And it really highlights that these two systems at least ( C64 and PC-88 ) both *did* resist the 86 PC onslaught, and resisted quite well. The other graph you posted on the PC-88 also shows a very similar pattern.... I was just focussed on the 1988 onwards part of the graphs, but the whole graph is pretty amazing.

And compared to CP/M computers in the US, I would argue both of these are 'late surviving' miracles... I'm looking for the reasons why.

You might think you made a joke here, but, no, this is *exactly* why nobody cared about CP/M. Unless you're running a dental office or accounting firm there was absolutely nothing interesting about it.

I didn't expect this point, but I did like it. I've always wondered about the BETA/VHS porn argument, though I did notice that Beta still put up a fight. I imagine there's some truth to your point.

If the PC98 was a hard-ware compatible upgrade for the PC88 then there must have been business use inbetween dithered anime girls, but I can't find numbers around that... Even if I did make a joke about the porn on the CP/M machines, the whole Text Only element *might* actually be a stronger elemenr than anything else... Amstrad PCWs had graphics. PC88s had graphics. Eastern European CP/M machines had graphics... US companies with CP/M history and graphics? Not as obvious, and not as usable.

I haven't heard the Compuserve porn premise before... Then GIF stands for Girls I'd ..... Fancy?

 

[cj7hawk]

Apple DOS is roughly as comprehensive as CP/M and TRS-DOS is significantly fancier than it, and both were pretty much written by one guy

I did want to ask you about this statement - You mention "Pretty Much".... and then "Paul Laughton and Randy Cook, respectively"

I found a fair bit from about Paul Laughton - and I can't find details of any other employees that contributed -It is reasonable to say he *did* write Apple DOS completely solo? I assume without any code input by Apple?

Likewise, Randy Cook seems to have a similar claim for TRSDOS, in which it appears there was no one else involved in it's writing at the time.

The reason I ask is because both seem to have been working on contract to their respective suppliers, though Randy Cook's story seems to have more of a bitter end to it. That aside, it is often said that "Gary Kildall" wrote CP/M - yet I've had it confirmed by people here on this forum that "Wrote" in that sense meant he was a major author and had others working for him write parts of it.

Do you have any details on whether Paul Laughton's and Randy Cook's work was entirely solo? Or whether they had other subcontractors/employees/contributors working for them who wrote sections?

Thanks
David

 

[vwestlife]

That's a pretty interesting graph. Makes me wonder what happened in 1987 to cause a resurgence for these machine that CP/M never took advantage of.

The Atari 7800, NES, and Sega Master System revitalized the 8-bit gaming market in the U.S., and the C64, which was often sold alongside them in stores, rode that wave.

(I know everyone says the NES was released in 1985, but that was just a test market of 100k units in NYC. The 7800 was the first to have a nationwide release in May 1986, followed by the NES and SMS in September.)

 

[DeltaDon]

Don't forget there was a Z80/CP/M card for the Apple II and that the II had graphics in its native mode and I guess was used by quite a few small businesses with CP/M programs. None of this mattered when the PC market took off.

 

[Agent Orange]

You know, in this whole discussion, it's not really about the hardware but it's the software and what could possibly run it. Data bases, spreadsheets, word processors, and games is what dictated the direction of the PC market in the early going. Today it's mostly has to do with what ever Microsoft decides. Microsoft dictates what features a motherboard must have and even the CPU's must conform to some degree. Also today's gaming industry must pay homage to and satisfy Microsoft with their code. Most everyone remembers when the government stepped in and took the Bell System apart be cause of their monopoly. Makes me think that in the back rooms in Redmond that they know that the clock is ticking and are just waiting for the axe to eventually fall.

 

[krebizfan]

Commodore had multiple Z80/CP/M addons planned for the various computer lines like the ill-fated CPM cartridge or boards for the PET and (proposed) CBM-II plus the inclusion on the C128. Many fairly capable computers were paired with a CP/M Z80 card if the Z80 secondary processor wasn't baked into the system. The sales of computers had been growing so fast that even an insignificant market share could still mean selling many thousands of cards and keep a company afloat.

 

[vwestlife]

Don't forget there was a Z80/CP/M card for the Apple II and that the II had graphics in its native mode and I guess was used by quite a few small businesses with CP/M programs.

Yeah... until like 1983 or so. During that time, Apple was pushing the III as their business computer, and basically let the II line languish. Once Apple came to their senses and introduced the IIe with 80-column text, more RAM, and ProDOS, that allowed a lot of business software to run natively on the II platform, instead of relying on a Z80 card and CP/M.

I've seen many II+'s with a Z80 card in them, but very few IIe's. And by the time the IIc came out, people were already looking at the 65C816 as the future of the Apple II line, not the Z80.

 

[alank2]

Most everyone remembers when the government stepped in and took the Bell System apart be cause of their monopoly. Makes me think that in the back rooms in Redmond that they know that the clock is ticking and are just waiting for the axe to eventually fall.

I'd like to think you are right, but the crazy in this country is way past splitting up large businesses because they have a monopoly. The government might threaten it, but honestly, there are massive numbers of businesses in this country that are massive and more or less in bed with our enemies. It is not what it used to be, not even close.

 

[krebizfan]

It was easy to split up AT&T into regional units plus long distance service since that was how AT&T was structured. The biggest effect of splitting MS into OS and application divisions would be to slow the deployment of any good ideas from Office into Windows for use by competitors. They would retain their dominant position through inertia.

 

[Eudimorphodon]

I found a fair bit from about Paul Laughton - and I can't find details of any other employees that contributed -It is reasonable to say he *did* write Apple DOS completely solo? I assume without any code input by Apple?

This is amply documented: here's a summary in CNET and here is his homepage. The TL;DR, is yes, Paul Laughton wrote the entire OS other than the driver for Woz's disk hardware (Woz and Bob Wigginton wrote that) in 35 days, single-handedly other than whatever little ancillary support he might have received from the other three employees at Shephardson Microsystems, the contracting firm he was hired through.

Apple paid $13,000 for the contract. That's more than the $90 Gnat Computing paid for the first-ever OEM CP/M distribution license in 1977, but less than the same license typically cost in 1978. Seems like a pretty fair deal.

Likewise, Randy Cook seems to have a similar claim for TRSDOS, in which it appears there was no one else involved in it's writing at the time.

Yes, Randy Cook wrote the original TRS-DOS single-handedly. The twisty part here is that Tandy didn't hire him directly; Tandy made a deal with Shugart to buy their disk drives with the proviso that Shugart provide a DOS to go with them. Shugart hired Cook, who wrote the thing from scratch in a couple months, provided Tandy with some drafts, and then an ugly game of telephone ensued when those buggy drafts ended up in customer hands and sparked a firestorm of complaints that ultimately spoiled the relationship between Cook and Tandy enough that Tandy elected to hire an in-house programmer to squash the bugs...

But, as the legend goes, Tandy was beaten to the punch by Apparat, an early third-party TRS-80 software vendor, who came up with their own patches to fix TRS-DOS 2.1 and sold the result as "NewDos". This set the precedent for the TRS-80 spawning a thriving ecosystem of third-party enhanced operating systems that lasted right up until the end of the platform. Cook himself wrote his own competing DOS called "VTOS", which had a lot of innovative and forward-looking features, but was kneecapped by also including an annoying copy production system; Randy was exceedingly bitter that Radio Shack (and Apparat) never paid him any per-copy royalties for his work (Tandy's defense being that it'd always been a flat fee, and Apparat's being they were just charging for patches to something Radio Shack gave away free with every disk drive purchase) and shot himself in the foot overreacting to it. VTOS *did*, however, succeed in turning into LDOS when Logical Systems Inc. licensed its source code and did away with the copy protection, ultimately becoming the last standing OS for the TRS-80 line. (In the form of LS-DOS for the Model 4.)

Anyway, I'm not sure what you're fishing for here. CP/M (and Apple DOS, and the original TRS-DOS) are not a lot of code (by today's standards), and they were written by trained professional programmers, it's not that crazy that they could be churned out by one guy in a couple months. (Randy Cook, for instance, had worked on several Minicomputer OSes before getting drafted for the TRS-DOS contract.) Writing an OS at least as-if-not-more complex than CP/M is pretty common task for computer science students these days, and you yourself have been busy writing your own CP/M clone.

I mean, sure, clearly *having* to write one is a significant enough effort that there's definitely logic in trying to license something that just needs a little customizing instead of doing it from scratch if your goal is to just sell a few computers, but both Apple and Tandy had already sold thousands of units before
deciding they needed to offer a disk drive, so it was perfectly logical for them to assume they'd have a big enough captive audience for the finished product to pay for the man hours to slap a custom one together.

I'm not suggesting these machines survived because they were CP/M machines... I'm wondering why these machines survived in Japan, Eastern Europe, Western Europe etc, but died out almost like they were hit by a meteorite in the US, then it spread rapidly. By 1990, the performance of x86 was enough, but early-mid 80s it's not like z80 based CP/M systems even tried to fight back. Companies either went directly PC or died out entirely. With the exception of Apple, who stuck to their architectures.

You keep repeating this premise, but have you ever considered that maybe this whole thread is based on a false idea here? Here's a few graphs from a well known Ars Technica article about computer market share.

Piece these graphs together in your head, and it's clear that 8-bit computers didn't just come to a screeching halt in the US much earlier than they did anywhere else. Do you remember earlier when I mentioned that 1985 was the year the aggregate sales of 16 bit computers exceeded the sales of 8-bit computers in the Japanese market? Look at the year 1985 in these graphs and you'll see that between the Apple II, the Commodore 64, and the little slivers of Atari 800 and "Other" left on the graph that was *also* the year that sales of 8 and 16 bit computers crossed over in the US. Yes, in subsequent years the share of the IBM PC continued a healthy growth pace while the Apple II and C64 stayed pretty much flat (and the last of the "others" died out) until around 1988, when Apple II sales started to crater, followed by the C64 around 1990, but... no, they didn't just "stop". In fact you really should leave Japan out of this because from what I can dig up (solid market share numbers for Japan seem to be really hard to come by) if anything the 8-bits dropped dead *faster* than they did in the US. The *diversity* of 8-bit machines certainly declined pretty fast post-1985 in the US, but that's because the smaller vendors largely went out of business or threw in the towel and started selling PCs instead because doing that was more economically efficient and more likely to succeed than wasting money on enhancing their obsolete products.

(Let's be clear, though, a few companies in the 1985-1987-ish timeframe *did* blow the money on churning out some massively enhanced "last hurrah" successors to classic old 8-bit machines in the US: Apple churned out the IIgs, Commodore gave us the 128, heck, even Tandy, who made it *really* clear that they'd decided the future was in MS-DOS, birthed the ridiculously-enhanced-compared-to-its-predecessor Color Computer 3. And these machines *did* manage to sell decently for a couple years to the most faithful true believers... but ultimately turned out to be dead ends because the end point of a massively enhanced computer from 1980 was pretty much where PC compatibles were starting from in 1987, and the growth path for PCs was already charted. These were essentially going away parties for old friends... just like the late models of the PC-88 were.)

As for Eastern Europe and whatnot, that's already been explained: These were markets that couldn't *afford* PC compatibles, not until they got *even cheaper*, so their curves lagged behind a few years. They dicked around with Sinclair clones until the mid-90's because they could buy one for literally nothing after Western Europeans didn't want them any more. There's no mystery here.

And it really highlights that these two systems at least ( C64 and PC-88 ) both *did* resist the 86 PC onslaught, and resisted quite well. The other graph you posted on the PC-88 also shows a very similar pattern.... I was just focussed on the 1988 onwards part of the graphs, but the whole graph is pretty amazing.

Just want to point out one more thing about the PC-88: you do understand that it was the upwards-compatible followup to the PC-8000 series, circa 1979, which was to the Japanese domestic market essentially the TRS-80 and the Apple II put together, right (in terms of both market share and cultural influence)? You made a swipe against Apple users as being essentially cultists, has it occurred to you that the reason the PC-88 had similar longevity to the Apple IIe in the United States could possibly be because NEC was the "Apple" of Japan?

 

[Eudimorphodon]

That's a pretty interesting graph. Makes me wonder what happened in 1987 to cause a resurgance for these machine that CP/M never took advantage of.

@vwestlife already gave you exactly the answer I was about to type out here. But let me expand upon it a little bit:

A thing you seem to really stubbornly be refusing to grasp is that there isn't one single monolithic "computer market"; from the day personal computers went from baffling kits that only wild-eyed HAM radio enthusiasts could assemble to something a mere mortal could walk into a store and buy the total market for computers has been segmenting into various niches, IE, "business computers", "home computers", "game computers", etc. Here's a thing you need to keep firmly in mind when you try to interpret those market share graphs I posted above: not only will one segment (tend to) have different requirements than another segment, those segments can be growing (or contracting) at different rates, and a computer that may have come to market in one segment might end up moving to another segment later on.

In those market share graphs I posted the ever-growing pile of blue (IE, PC compatibles) at the bottom of the graph was supercharged in large part not because individuals were buying PC compatibles for home use (although that became more and more normalized as time went on), but because businesses were automating office work en-masse through the 1980's. Sure, this is a thing that started in the 1970's and the first wave of it *did* involve CP/M machines, but well into the 1980's it was perfectly normal to see the receptionist sitting in front of a typewriter instead of a computer or a business entering every transaction into one of those big double-entry bookkeeping journals. And in addition to the hundreds of thousands of small businesses that were buying their first computers in the latter half of the 1980's you also had massive Fortune 500 companies buying millions of PCs to upgrade and replace older office automation systems involving tabulation machines, IBM 3270 terminals, etc, etc. These numbers showing 8-bit computers just "dropping dead" in the US in the mid-1980's are warped by all these business buying "business machines", and a "business machine" meant an IBM compatible by around 1984 or so.

(Again, I'm sure you're dying to say "but why didn't CP/M, specifically, compete!?!", you've already had the answer dozens of times over: The installed base of CP/M was never that big in the first place, PCs didn't cost any more than a business-class CP/M computer, and the benefits of disk and program compatibility minus the CP/M faffing around made it a complete no-brainer for the early adopters to migrate.)

... anyway, like I said, 8-bit machines didn't drop dead all at once in the home market (which since you keep bringing up toys like Sinclair Spectrums and obsessing over the number of games per year that came out for them must be what you’re talking about). A phenomenon that did happen was a few machines that had been priced more like "business computers", IE, the Apple IIe, when they were introduced ended up getting elbowed out of the "business" niche and spent their last few years sharing the "home" niche with computers that had been born into it like the C64, and the C64 itself, near the end when "home computer" really did start meaning just a low-spec affordable PC, still managed to eek out a bit more life by niche-shifting into the "game computer/console" market. Which, again, exactly describes the career of the Sinclair Spectrum; it was *never* a "business class" computer, ever, and I'd argue it was in fact aimed so low it was almost born directly into the "game computer" niche instead of falling into it.(*)

Basically the thing you're looking for here, some real continuum between CP/M business computers of the late 1970's and the last crop of Z80 game computers that *could* run CP/M (either directly or with hacking) doesn't really exist. Business CP/M, the domain of teletypewriters and Wordstar, died a well deserved death in the early 1980's because it sucked. CP/M being a "thing" on late 80's game computers was essentially an exercise in zombie computing.

(* by contrast the PC-88 platform appears to be a legit case of a once-business-class machine dropping all the way down through “home computer” to “game computer” over its life. I guess if we count educational use in the “Game” category the Apple II also pulled off the full sweep; the education bulk price of an Apple IIe near the end was close to ZX spectrum territory.)

 

[cj7hawk]

... anyway, like I said, 8-bit machines didn't drop dead all at once in the home market (which since you keep bringing up toys like Sinclair Spectrums and obsessing over the number of games per year that came out for them must be what you’re talking about). A phenomenon that did happen was a few machines that had been priced more like "business computers", IE, the Apple IIe, when they were introduced ended up getting elbowed out of the "business" niche and spent their last few years sharing the "home" niche with computers that had been born into it like the C64, and the C64 itself, near the end when "home computer" really did start meaning just a low-spec affordable PC, still managed to eek out a bit more life by niche-shifting into the "game computer/console" market. Which, again, exactly describes the career of the Sinclair Spectrum; it was *never* a "business class" computer, ever, and I'd argue it was in fact aimed so low it was almost born directly into the "game computer" niche instead of falling into it.(*)

I've never disagreed that the other machines of the era, all the ones you've mentioned, really hung in there well into the PC era... And for various reasons, including being the Apple of their respective markets ( for both the NEC and the Amstrad PCWs ) a number of machines did thrive and had several "last hurrah" releases. And non-CP/M lines definitely went further (eg, Amiga, ST ) and of course Apple survived, although any one of it's platforms would have died out but the line of machines continued.

I was just following the logic that this didn't appear* to have happened in the US with CP/M machines.

* You note I could be incorrect in my observation, and that may be, but I still don't know of any US made CP/M machines that tried to fight MS-DOS's inevitable march. I think at this point it's reasonable to conclude there weren't any. Many interesting reasons have been provided. But it doesn't appear any hung around like the other machines of the era.

As for games sales? Or software sales at all? It's not a perfect indicator, but it does give an indication into the market, whether the machines are selling or not. The graphs you provided show a really good example of the popularity of the platform and the value of producing software for it across it's lifetime, along with illustrating when it died.... And since you used the term, I guess a zombie is any use of the platform post-death.

 

[cj7hawk]

Anyway, I'm not sure what you're fishing for here. CP/M (and Apple DOS, and the original TRS-DOS) are not a lot of code (by today's standards), and they were written by trained professional programmers, it's not that crazy that they could be churned out by one guy in a couple months. (Randy Cook, for instance, had worked on several Minicomputer OSes before getting drafted for the TRS-DOS contract.) Writing an OS at least as-if-not-more complex than CP/M is pretty common task for computer science students these days, and you yourself have been busy writing your own CP/M clone.

It was just a question about some historical events you mentioned that I wasn't aware of, that I would have assumed were approached differently. Your answer was helpful and appreciated and learning more about other efforts like this is something I became interested in after writing a clone OS. Understanding the mechanics of other people's architectures is enjoyable I find examples such as this as enjoyable as learning about machine architectures I didn't know about.

I too wrote the basics of an OS when a student, but always in high level languages. I found that It's a very different thing to write the entire program in Assembly. Tracking down other examples is something I've been doing.

 

[krebizfan]

I've never disagreed that the other machines of the era, all the ones you've mentioned, really hung in there well into the PC era... And for various reasons, including being the Apple of their respective markets ( for both the NEC and the Amstrad PCWs ) a number of machines did thrive and had several "last hurrah" releases. And non-CP/M lines definitely went further (eg, Amiga, ST ) and of course Apple survived, although any one of it's platforms would have died out but the line of machines continued.

I was just following the logic that this didn't appear* to have happened in the US with CP/M machines.

* You note I could be incorrect in my observation, and that may be, but I still don't know of any US made CP/M machines that tried to fight MS-DOS's inevitable march. I think at this point it's reasonable to conclude there weren't any. Many interesting reasons have been provided. But it doesn't appear any hung around like the other machines of the era.

As for games sales? Or software sales at all? It's not a perfect indicator, but it does give an indication into the market, whether the machines are selling or not. The graphs you provided show a really good example of the popularity of the platform and the value of producing software for it across it's lifetime, along with illustrating when it died.... And since you used the term, I guess a zombie is any use of the platform post-death.

Lots of US based CP/M systems tried to forestall the IBM PC's takeover. Look at the marketing for Compupro and Kaypro for examples. Even the DEC Rainbow pushed the idea of running CP/M on the Z80 instead of using DOS with 8086. Of course, the time lag in getting designs to market meant that the lack of DOS applications was alleviated long before the advertisements were released.