IBM PC: The complete history, part 2

The real victor was Microsoft, which built an empire on the back of a shadily acquired MS-DOS.

Read the whole story

 

[TheThirdDictor]

Eh. I agree with you that the way Microsoft originally acquired 86-DOS was shady (claiming they had licensees to buy it and offering to pay SCP $10k for each license sold). But that's not the nail that sealed 86-DOS's coffin.

The big screw up is when Rod Brock started shopping 86-DOS around. Not only did he entertain Microsoft as a client, he agreed to turn over complete ownership to them and himself settle for an exclusive license. Remember, originally this is the way he wanted to license to the prospective customer in the first place. He had no intent of surrendering complete control and accepting only a license in return. Yet, for a measly $50k he did.

Fool me once, shame on you, fool me twice, shame on me. Obviously Rod Brock didn't have all the information available, but Microsoft already deceived him once. $50k is a lot when your business isn't doing well, especially in the early 80s... but come on. He should have known better after the first deception.

The article mentions how this was common in the era, how computer programmers didn't know how to be businessmen. Fine. There's a difference between not being proficient in business and intentionally doing business with a company you know deceived you in the past.

Right. If Bill Gates is willing to give you $50,000? You *know* it's worth a lot more. But to leverage that, you have to be in a position where $50k isn't going to save your business. You just have to be capitalized well enough to laugh in his face at the table, and say "Come back when there is another digit in the offer."

Nothing wrong with doing business with somebody who has screwed you before..happens all the time in business. But after the first screwing? You simply *have* to be smart enough to screw back.

 

[Jeremy Reimer]

I hope ars follows up with stories about some of MS's nefarious deeds that followed the
introduction of the IBM PC to undermine competitor applications running on MSDOS.

The whole IBM vs MS war over OS/2 and Windows is a great story arc to continue the
adventure.

Well, if you really want to read more about Microsoft, IBM, and OS/2 without having to wait, you can read this Ars article.

 

[operagost]

Novell finally bought Digital in 1991, allowing Kildall to retire a millionaire.

Why does the author drop "Research" from Digital Research's name in many places (including in the original article)? It actually bothers me and is incorrect. Digital Equipment Corporation, aka DEC aka Digital (a trademark), was a completely different company purchased by Compaq in 1998.

Even the Wikipedia article has a "Not to be confused with Digital Research" prologue.

Ironic that Digital Research came out with DR-DOS many years later (as Windows was coming into vogue).

DR started out as "Intergalactic Digital Research", as I recall.

I guess Kildall later came to the realization that was a bit too ambitious a moniker, even in the forward-looking computer industry.

 

[Leguleius1]

I'm not an expert on common law but I do believe that parties' expectations don't count for much when dealing with contracts and contract breach. What I can tell you is that in continental European legal systems that mental reservation could come back to kick Bill Gates in the butt.

....

I think you're missing just a couple of things:

First, the QDOS developer may have fallen into the "rose colored glasses" pitfall. Gates could say, with perfect honesty, "we see this taking off, and would not be surprised to see 500,000 machines sold in the first year." Gates would be referring to a SINGLE customer (IBM) and thus only a single license payment, but the developer would see that through rose tinted lenses as potentially more.

Second, remember that the province of QDOS has been called into questions in terms of whether any lines were crossed relative to protected information from CP/M. The developer may have had more reservations about that issue than we see here decades later, and that may have played into the deal that was made.

One thing I know, is that the case books are filled with all sorts of interesting lawsuits over tech licensing from the early age of computers. If there was some sort of fraud or other "legally actionable" shady behavior that could be proven, the sheer amount of money involved suggests it WOULD have been litigated. The fact it wasn't suggests maybe it was just a "bad deal."* Which goes back to my post in this thread about the time Microsoft itself made a "bad deal" of the same sort relative to the Basic used in Commodore computers.

*Edited to add: ESPECIALLY since MS and the DOS developer wound up in litigation later anyway.

 

[sigma8]

Great article! I hope I have time later to read the parts I had to skim over for now. I jumped into the field a bit after this, more 1985ish.. and of course when I did, I was just a kid so wasn't paying any attention to industry drama beyond getting Decathlon and Starflight to work.

I know the whole "Microsoft snatching up DOS" story makes IBM sound like a huge victim and like MS pulled great shenanigan, but to be fair, MS didn't rest on its laurels, and IBM didn't really suffer for many years after this happened...and I'd say their failures and successes can't really be attributed to this moment, either.

MS, while never bleeding edge, kept DOS relevant and frequently updated, and IBM continued to be the de facto hardware platform -- neither DOS nor Windows 1/2/3 attempted to change any of that.

I think the stupidity from IBM came from MCA and OS/2--both being horrible executions of relatively good ideas, and their losses were gains for Compaq/Dell/Third-parties (hardware) and Microsoft (software).

 

[WaveRunner]

I wonder if Patterson still works at Microsoft... don't know the exact details but my understanding is Bill squared the deal when he hired him with options worth north of a million. Unfortunately I don't think Patterson held on to them for very long. Bill Gates is very competitive but wronging Patterson is obviously not part of his morals.

 

[Nerdferklel]

That was a very engrossing read. Thanks Ars!

 

[NdYAG]

Great articles. One observation that I've had for at least the 30 subsequent years after the IBM PC was that no matter how powerful the PC's got, the cold boot times stayed annoyingly similar.

Things are different now with sleep functions and SSD's, but I do miss the slow memory check of the early PC's. I started each day with seeing something successful happen.

 

[raxx7]

I wonder if Patterson still works at Microsoft... don't know the exact details but my understanding is Bill squared the deal when he hired him with options worth north of a million. Unfortunately I don't think Patterson held on to them for very long. Bill Gates is very competitive but wronging Patterson is obviously not part of his morals.

According to Wikipedia, Tim Paterson has worked for Microsoft for 3 periods of time, the last ended in 1998.

 

[raxx7]

Great articles. One observation that I've had for at least the 30 subsequent years after the IBM PC was that no matter how powerful the PC's got, the cold boot times stayed annoyingly similar.

Things are different now with sleep functions and SSD's, but I do miss the slow memory check of the early PC's. I started each day with seeing something successful happen.

I think most BIOS still have that as an option.

 

[greevar]

If one is going to condemn QDOS for being a rip-off of CP/M, then something else must be considered.

A smoking gun by the name of PIP (Peripheral Interchange Program).

Before there was CP/M, there was OS/8, and I've been told there was also an operating system for the PDP-11 to which CP/M was very similar indeed.

"Great artists steal, but not from me..."

 

[beausoleil]

I think the stupidity from IBM came from MCA and OS/2--both being horrible executions of relatively good ideas, and their losses were gains for Compaq/Dell/Third-parties (hardware) and Microsoft (software).

I'll agree that the Micro Channel Architecture was really mis-handled - not being backward compatible with the PC XT and AT bus architecture.

But OS/2 was really, really nice - once Microsoft got out of the picture after screwing IBM by developing NT "on the sly".

Once OS/2 Warp shipped with built-in IP networking and the next gen GUI, it was really, really sweet. Fully 32-bit, it also included a working Windows 16-bit subsystem which could run most Windows apps natively. And the HPFS file system included most all the nice features that would later be included in NT's NTFS...

 

[raxx7]

I think the stupidity from IBM came from MCA and OS/2--both being horrible executions of relatively good ideas, and their losses were gains for Compaq/Dell/Third-parties (hardware) and Microsoft (software).

I'll agree that the Micro Channel Architecture was really mis-handled - not being backward compatible with the PC XT and AT bus architecture.

While the MCA slots were not backward compatible with ISA cards, you could mix up MCA and ISA slots in the same mainboard, pretty much the same way we later mixed up ISA and PCI or nowadays we mix up PCI and PCIe (well, if you search hard enough, you might just find a computer with ISA, PCI and PCIe).


<edit>Not even hard actually. Behold: https://www.amazon.com/MB-P4BWA-Industr ... B004HLOQH6></edit>

The problems with MCA were (a) licensing and (b) some technical aspects.
Eg, the plug n' play model of MCA was... f**ked up.

But OS/2 was really, really nice - once Microsoft got out of the picture after screwing IBM by developing NT "on the sly".

Once OS/2 Warp shipped with built-in IP networking and the next gen GUI, it was really, really sweet. Fully 32-bit, it also included a working Windows 16-bit subsystem which could run most Windows apps natively. And the HPFS file system included most all the nice features that would later be included in NT's NTFS...

It was also heavy on resources and had problems with backward compatibility with DOS applications.
A problem that Windows NT also suffered.
In hindsight, the Windows 9x hybrid clusterf**k was the only path that the market was willing to accept.

 

[kalzekdor]

Second, remember that the province of QDOS has been called into questions in terms of whether any lines were crossed relative to protected information from CP/M. The developer may have had more reservations about that issue than we see here decades later, and that may have played into the deal that was made.

Apologies in advance, but I believe the word you're looking for is "provenance".






......I'll see myself out.

 

[Swimminginurea]

Novell finally bought Digital in 1991, allowing Kildall to retire a millionaire.

Why does the author drop "Research" from Digital Research's name in many places (including in the original article)? It actually bothers me and is incorrect. Digital Equipment Corporation, aka DEC aka Digital (a trademark), was a completely different company purchased by Compaq in 1998.

Even the Wikipedia article has a "Not to be confused with Digital Research" prologue.

Ironic that Digital Research came out with DR-DOS many years later (as Windows was coming into vogue).

DR-DOS was the best DOS I have ever used. I was sad when I finally made the leap into Windows.

 

[NinjaNerd56]

I was recruited on campus in 1982....by Radio Shack...to run a Radio Shack Computer Center after college.

Took the gig and made a shit ton of $$ selling TRS-80s. Made Club in 84 and got a week in Hawaii. Resigned a month after I got home and opened an IBM/Compaq retail store. Never looked back.

Interestingly, I figured out very early that selling $200 Color Computers was not the path to riches.

I mostly sold Model 16s running MS-XENIX and Realworld COBOL accounting software. Average system price was ~$30K.

Most folks may not know or recall that the 16/6000 machines had a Z80 CPU daughter card along with a Motorola 68000 'main' chip.

If you booted TRS-DOS, the Z-80 owned the box. Boot XENIX, and the Z-80 became a serial I/O manager.

Fun times....

Once wrote 30K lines of BASIC to create RISK on the CoCo. It was as good as anything that exists today; the CPU was actually a pretty good opponent.

Offered it to Parker Brothers, who passed, and then insult to injury, wouldn't let me market it. The hardcopy is long gone, but I still have a CASSETTE TAPE from a CCR-81 tape deck with the code.

*SIGH*

 

[odaiwai]

Awesome read.
Now I can see where Halt and Catch Fire comes from....

https://smile.amazon.com/Soul-New-Machi ... ew+machine

1981.

 

[SilverSee]

So how exactly was it shady? The guy had a product but couldn't figure out what to do with it. Gates saw an opportunity and made it into something. If Steve Jobs had done the same everyone would be saying what a great mind he had.

I sort of agree. It is at least certain that Microsoft was contractually bound to secrecy by IBM so they could not have revealed who the license was for or been more forthcoming about their plans.

 

[daemonios]

I'm not an expert on common law but I do believe that parties' expectations don't count for much when dealing with contracts and contract breach. What I can tell you is that in continental European legal systems that mental reservation could come back to kick Bill Gates in the butt.

....

I think you're missing just a couple of things:

First, the QDOS developer may have fallen into the "rose colored glasses" pitfall. Gates could say, with perfect honesty, "we see this taking off, and would not be surprised to see 500,000 machines sold in the first year." Gates would be referring to a SINGLE customer (IBM) and thus only a single license payment, but the developer would see that through rose tinted lenses as potentially more.

Second, remember that the province of QDOS has been called into questions in terms of whether any lines were crossed relative to protected information from CP/M. The developer may have had more reservations about that issue than we see here decades later, and that may have played into the deal that was made.

One thing I know, is that the case books are filled with all sorts of interesting lawsuits over tech licensing from the early age of computers. If there was some sort of fraud or other "legally actionable" shady behavior that could be proven, the sheer amount of money involved suggests it WOULD have been litigated. The fact it wasn't suggests maybe it was just a "bad deal."* Which goes back to my post in this thread about the time Microsoft itself made a "bad deal" of the same sort relative to the Basic used in Commodore computers.

*Edited to add: ESPECIALLY since MS and the DOS developer wound up in litigation later anyway.

I'm sure I'm missing out a LOT of things. Which is why I put a disclaimer right there on my post saying it was just a scenario to show how the deal *could* be construed as less-than-honest.

 

[daemonios]

I'm not an expert on common law but I do believe that parties' expectations don't count for much when dealing with contracts and contract breach. What I can tell you is that in continental European legal systems that mental reservation could come back to kick Bill Gates in the butt.

....

I think you're missing just a couple of things:

First, the QDOS developer may have fallen into the "rose colored glasses" pitfall. Gates could say, with perfect honesty, "we see this taking off, and would not be surprised to see 500,000 machines sold in the first year." Gates would be referring to a SINGLE customer (IBM) and thus only a single license payment, but the developer would see that through rose tinted lenses as potentially more.

Second, remember that the province of QDOS has been called into questions in terms of whether any lines were crossed relative to protected information from CP/M. The developer may have had more reservations about that issue than we see here decades later, and that may have played into the deal that was made.

One thing I know, is that the case books are filled with all sorts of interesting lawsuits over tech licensing from the early age of computers. If there was some sort of fraud or other "legally actionable" shady behavior that could be proven, the sheer amount of money involved suggests it WOULD have been litigated. The fact it wasn't suggests maybe it was just a "bad deal."* Which goes back to my post in this thread about the time Microsoft itself made a "bad deal" of the same sort relative to the Basic used in Commodore computers.

*Edited to add: ESPECIALLY since MS and the DOS developer wound up in litigation later anyway.

I'm sure I'm missing out a LOT of things. Which is why I put a disclaimer right there on my post saying it was just a scenario to show how the deal *could* be construed as less-than-honest:

This is just to explain why the deal might be seen as shady, but note that I don't have all the details, and the parties came to an agreement in the end anyway, so that point is kind of moot.

 

[rbaillie]

Interesting story, thanks for writing it!

One of my favorite books about economic-type issues is, "Billionaires' Ball" by Linda McQuaig and Neil Brooks. Among other things, the book talks about how Bill Gates benefitted from the types of things discussed above, in addition to being born into a wealthy family, and being born into a society that offers patent protection, the rule of law, etc.

 

[KurtisMayfield]

Not one mention of why IBM bet on Bill Gates.. the family connections!

His mother knew the CEO of IBM from working together at the United Way

Without this, none of this happens.

 

[f00barbob]

Haven't encountered any 5150s in recent history, but the only real problems I've encountered with older ones are spontaneously-exploding tantalum capacitors. I learned from some old Hyundai 286 machine that first power on test after long term storage is best conducted with the case closed up... almost got nailed in the eye.

 

[Hapticz]

for a kid that didn't finish harvard, gates reminds me of another harvard man that started another two faced operation with sly methods. maybe its something in the cambridge water supply.

when faced with actually having to find something to live off of, these types act no classier than the mob, or politicians.

thanks bill, you helped me present my opinions via a well rooted ms system!

 

[krebizfan]

Eh. I agree with you that the way Microsoft originally acquired 86-DOS was shady (claiming they had licensees to buy it and offering to pay SCP $10k for each license sold). But that's not the nail that sealed 86-DOS's coffin.

The big screw up is when Rod Brock started shopping 86-DOS around. Not only did he entertain Microsoft as a client, he agreed to turn over complete ownership to them and himself settle for an exclusive license. Remember, originally this is the way he wanted to license to the prospective customer in the first place. He had no intent of surrendering complete control and accepting only a license in return. Yet, for a measly $50k he did.

Fool me once, shame on you, fool me twice, shame on me. Obviously Rod Brock didn't have all the information available, but Microsoft already deceived him once. $50k is a lot when your business isn't doing well, especially in the early 80s... but come on. He should have known better after the first deception.

The article mentions how this was common in the era, how computer programmers didn't know how to be businessmen. Fine. There's a difference between not being proficient in business and intentionally doing business with a company you know deceived you in the past.

Right. If Bill Gates is willing to give you $50,000? You *know* it's worth a lot more. But to leverage that, you have to be in a position where $50k isn't going to save your business. You just have to be capitalized well enough to laugh in his face at the table, and say "Come back when there is another digit in the offer."

Nothing wrong with doing business with somebody who has screwed you before..happens all the time in business. But after the first screwing? You simply *have* to be smart enough to screw back.

Except it was fairly clear at the time that MS was overpaying. DRI charged IMSAI $25,000 for a complete license to CP/M. MS paid double that for an OS everyone expected to be pushed aside as soon as CP/M-86 was completed. MS offered to let IBM purchase the license directly from SCP because it was clearly going to be a money loser and only did the license to save the rest of the PC contract. The expected future in 1980 was considerably different from what actually happened.

 

[Jimmy the Geek]

This is the concluding part of the surprisingly interesting history of the IBM PC.

I hope this is an erroneous statement. There should absolutely be a Part 3 about the BIOS and its successful reverse-engineering.

If IBM had been successful in keeping fully-compatible PCs out of the market, there's no telling how drastically different the industry would be today. Maybe IBM would have been able to leverage ownership of the BIOS into widespread adoption of the Micro Channel bus, and thus slowly moved all of the PC's systems into proprietary status. Without the windfalls of an exploding PC market, it's possible that both Microsoft and Intel would have had far less success and exerted far less influence. Which might have opened some doors for Steve Jobs and Company...

A world where Phoenix doesn't reverse-engineer the BIOS could be a world where Apple is the dominant name in home computing, while IBM is the higher-priced boutique product. Both use Motorola CPUs, but perpetual outsider Intel keeps trying to make inroads...

 

[Madlyb]

Loving this series, keep it coming.

 

[Danathar]

I often wonder what today would look like had Commodore not been run by a bunch of management idiots.

 

[TechMatters]

Brock signed the deal.

Headline:
The real victor was Microsoft, which built an empire on the back of a shadily acquired MS-DOS.

Story Details:
Brock was uncertain, but he really did need the money, and he didn't know what to do with 86-DOS himself anyway…
He signed the agreement, making Microsoft the sole owner of 86-DOS—or, as it was immediately renamed, MS-DOS.

Brock signed. Blaming Gates/Microsoft for making a smart decision is hilariously ludicrous.

 

[jkp2505]

I love the "legendary Gates shiftiness" tag you've given him.

Let's not forget it was the 'great' Steve Jobs who screwed his own friend out of more than $2000. Wozniak did all the hard work on creating the Breakout game for Atari then Jobs pocketed most of the cash and lied about how much they paid. Sounds pretty shifty to me.

 

[colinstu]

What a great series. This could be turned into a video documentary easily I think... just reading through it I'm imagining the interviews and craftily put-together dramatizations etc.

Will you do another write-up on IBM making OS/2 and trying to get back at MS?

 

[Jeff S]

An excellent read, thanks.

On the horizon, there are new architectures that might give us persistent RAM. You won't need a disk anymore - disk and RAM will be combined. That will require a major change in how operating systems and programs handle the hardware.

Will it require a major change? It'll require some changes, I don't know how major though . . .

It's still convenient to divide up storage into a filesystem with directories, names, ownership, permissions/ACLs. You'll still want a filesystem for all that.

It's also still true that you'll want an area of 'working memory' that's not really persisted. Assuming they're one in the same, you would probably just partition the memory up, similar to how a hard drive might currently be partitioned, with one working RAM partition, and the rest as filesystem partitions.

Conceivably, some type of apps (like an RDBMS) might skip using the filesystem, and have their own persistent partitions of memory that they just work directly with as persistent RAM devices (I think that, is it Oracle? Has used their own HDD partitions for many years, skipping the filesystem, so that's not without precedent even with HDDs).

Operating systems like Linux (and I think Windows) have already had a concept of MMIO (Memory-mapped I/O) where the file on disk is mapped to memory addresses and the file is essentially mirrored in RAM while it's open, with any changes to the version in RAM being automatically written out to disk. With persistent RAM, the API used by the app could be exactly the same, and the OS just knows that since it's persistent RAM, there's no benefit to copying it to memory first before the app uses it - just hand the app the addresses of the files (although, you still might not want to do that, because of stuff like journaling, which provides transaction-like updates to the files to help reduce corruption in case of reboot/lost power).

 

[DKlimax]

I hope ars follows up with stories about some of MS's nefarious deeds that followed the
introduction of the IBM PC to undermine competitor applications running on MSDOS.

The whole IBM vs MS war over OS/2 and Windows is a great story arc to continue the
adventure.

The funny thing is Gates is doing a lot a truly good things these days. Perhaps enough
to rehabilitate his name for posterity.

Are you referring to nonsense like "DOS ain't done till Lotus doesn't run"?

 

[Jsl]

Great read. I got my first experiences with PCs when I was around 5-6 years old in the late 1980s. My dad was using it to write his thesis and it came with some games preinstalled.

I have a question to somebody who was around a little earlier, or just knows their PC history better: The article seems to refer to DOS (which I know as an operating system) on equal terms with BASIC and Pascal (which I know as programming languages). Why is that? Was the distinction somehow more blurry in the early days of personal computing?

 

[macatari]

Articles like this make me so happy. More like this, please.

Ars is just republishing this. For more great articles like this, just go directly to the source, The Digital Antiquarian:

http://www.filfre.net

 

[JimmiG]

This is the concluding part of the surprisingly interesting history of the IBM PC.

I hope this is an erroneous statement. There should absolutely be a Part 3 about the BIOS and its successful reverse-engineering.

If IBM had been successful in keeping fully-compatible PCs out of the market, there's no telling how drastically different the industry would be today. Maybe IBM would have been able to leverage ownership of the BIOS into widespread adoption of the Micro Channel bus, and thus slowly moved all of the PC's systems into proprietary status. Without the windfalls of an exploding PC market, it's possible that both Microsoft and Intel would have had far less success and exerted far less influence. Which might have opened some doors for Steve Jobs and Company...

A world where Phoenix doesn't reverse-engineer the BIOS could be a world where Apple is the dominant name in home computing, while IBM is the higher-priced boutique product. Both use Motorola CPUs, but perpetual outsider Intel keeps trying to make inroads...

This is true. It wasn't the IBM PC, but cheap, reverse engineered IBM clones, "compatibles", that took over the home PC market and wiped out Amiga, Atari, Acorn, and very nearly, the Mac. Considering Amigas, Ataris and Macs all used Motorola CPUs, the CPU landscape could indeed have looked very different if the compatibles hadn't seen such widespread adoption.

 

[graylshaped]

My parents bought an Epson QX10 running Valdocs in the early 80's, and I remember driving home to use it for undergraduate papers in lieu of the Brother electric typewriter that had been my high school graduation gift. When I finished my undergraduate, they gave me a Kaypro transportable. They told me the salesman had shown them the new IBM PC's, which were slightly more expensive. "What do you get for the extra cost?" they asked. He thought, then slapped it and said "It's built like a tank. You could throw this out of a car and it would probably be ok."

They asked me "Are you planning on throwing this out of a car? If so, we'll take the Kaypro back and exchange it."

I used that Kaypro probably for five years, through graduate school and then some, until the monitor gave up.

Good old CP/M.

 

[Jeff S]

Brock signed the deal.

Headline:
The real victor was Microsoft, which built an empire on the back of a shadily acquired MS-DOS.

Story Details:
Brock was uncertain, but he really did need the money, and he didn't know what to do with 86-DOS himself anyway…
He signed the agreement, making Microsoft the sole owner of 86-DOS—or, as it was immediately renamed, MS-DOS.

Brock signed. Blaming Gates/Microsoft for making a smart decision is hilariously ludicrous.

Yeah, Brock did sign. He really should have consulted a lawyer, but probably didn't have even enough money for that. He made several mistakes:

Licensing that allowed sub-licensing without a royalty on each copy. Apparently Microsoft's contract with Brock defined an OEM like IBM (but not IBM by name) as a 'customer' who could be licensed, and make as many copies as they wanted.

In truth, the customers should have been defined as everyone who bought a computer that included a copy.

He could have asked for a dollar a copy and made far more than he made. But, he didn't license it by copy. He licensed it by OEM, which never probably would have made him that much money anyhow.

He let the developer of the software go work for his only customer, meaning he had no one to maintain it.

In the end, DOS probably wouldn't have evolved into what it needed to be anyhow without Microsoft's acquisition, and would have held back the entire PC industry.

Most of the value in DOS wasn't there in version 1.0. Microsoft's programmers built most of the actual value that was in DOS which came in the later versions like 2, 3, 4, and 4.1.

 

[krebizfan]

Great read. I got my first experiences with PCs when I was around 5-6 years old in the late 1980s. My dad was using it to write his thesis and it came with some games preinstalled.

I have a question to somebody who was around a little earlier, or just knows their PC history better: The article seems to refer to DOS (which I know as an operating system) on equal terms with BASIC and Pascal (which I know as programming languages). Why is that? Was the distinction somehow more blurry in the early days of personal computing?

UCSD P-System which included Pascal and FORTRAN as languages was a complete operating system that wrote to disks in its own unique format with an integrated development environment. Very popular in academia at the time.

If PC purchasers hadn't bought floppy drives with it, IBM was prepared to push cassette tapes. Typing Tutor was advertised in cassette form though that never shipped. ROM BASIC was popular in home computers and IBM had done the ROM BASIC plus tape successfully with the 5100.

IBM did not know how the PC would be used so planned for the most likely scenarios to give people reasons to buy an unproven system. Want an OS that permits easy porting from CP/M; IBM was ready. Want a ported minicomputer programming system; IBM had that. Want self-booting software; documentation was available to everybody. Want to be cheap and only use cassettes; visit a Radio Shack for the cable and go for it.

 

[real mikeb_60]

Excellent set of articles. As others have noted, a couple more episodes are probably in order ... someday.

Should have commented on part 1, but there was another IBM microcomputer pre-PC: Displaywriter. IIRC, it was even a full 8086 rather than 8088. But it was a dedicated word processor (and actually quite a good one if you sprang for the high-end packages). Occasionally used one in a long-ago job. According to Wikipedia, CP/M-86 was available for it though I doubt many people exercised the (expensive) option.

My wife worked on a System/23 for a while, setting up and running the bookkeeping for a contractor.

"DOS" back in the early days, except for CP/M, usually meant a layer on top of the system ROM (not yet called BIOS) that implemented the functions needed to read/write/build the filesystem on disk. Apple, Radio Shack, and Commodore all had those before IBM, and they worked essentially the same way: the ROM had hooks to call the DOS functions, but they did nothing unless a suitable disk was present. PCs at the time were really BASIC machines, running directly from ROM (which was about as fast as RAM at the time). CP/M was different, in that it didn't use the system ROM and booted entirely from disk, as did later versions of MS-DOS.

As for the "email" system - since MS was working for IBM, they probably had access to an IBM mainframe running VM/CMS. By the late 1970s, RSCS had evolved to where exchanging files between VMs and their mainframe hosts was relatively routine, and "mail" using the protocol was a trivial addition.