DR DOS was a third party clone of MS DOS that developed a well-deserved niche following in the late 1980s and early 1990s. Even though it was first released in the mid-1980s, its copyright date of 1976 suggests some historical significance. Indeed, it was more than just any clone of MS-DOS. Arguably it was both a clone and a predecessor.

The ill-fated deal with IBM

CP/M-86

After IBM priced CP/M-86 at $240, Digital Research sold CP/M-86 itself as an alternative operating system in 1983 but was unsuccessful.

DR DOS is a direct descendant of CP/M, the most popular operating system of the 1970s that ran on Intel 8080 and Zilog Z-80 CPUs. IBM famously wanted to sell a rebranded CP/M with its original IBM PC from 1981. But Digital Research found its terms to be too one-sided, and they were not able to reach a middle ground.

Microsoft knew of a CP/M clone being produced by another smaller company in Seattle. Microsoft licensed, and later acquired that product and rebranded it as MS-DOS. It also supplied it to IBM under license, allowing them to resell it as PC DOS.

And the rest is history. The way far too many people tell the story, Gary Kildall kept IBM waiting for hours while he was joyriding in his airplane, the genius Bill Gates gave IBM the respect they deserved, Gates became a billionaire, Kildall died in obscurity, and capitalism worked.

Besides being a massive oversimplification, Digital Research didn’t just go quietly into the night. Not by a long shot.

The CP/M-86 story

There was a good reason why Digital Research had been dragging its feet on producing an 8086 version of CP/M. They were working on something better. The 8080 processor wasn’t really cut out for multitasking, but the 8086 processor probably could. So they were working on a CP/M-like operating system with multitasking that ran on the 8086 and 8088 CPUs at the time when IBM came calling.

Digital Research ended up producing an 8086 port for the IBM PC and other 8086 and 8088-based computers. They called it CP/M-86. IBM even agreed to sell it as an option to avoid a lawsuit. PC DOS was a very close clone of CP/M. Kildall believed it infringed his copyrights. Under the terms of their settlement, IBM hired Digital Research to write its ROM BIOS. IBM also bought 2,000 copies of CP/M-86 to offer as an alternative operating system. Digital Research agreed not to sue either IBM or Microsoft.

But then IBM sold PC DOS for $40 while charging $240 for CP/M-86. As you might guess, CP/M-86 did not garner significant market share. Digital Research sold CP/M-86 itself at a competitive price in the 1982-1983 timeframe, with the packaging mentioning the IBM PC and XT by name. But even if they could get the product into all of the IBM dealers, IBM PC DOS in IBM packaging priced at $40 seemed like a safer, wiser option.

Building DOS compatibility into CP/M

But since MS-DOS was little more than a clone of CP/M running on a slightly newer and similar but not quite compatible CPU, there was little reason that Digital Research couldn’t turn around and make CP/M more MS-DOS compatible.

Remember those products with multitasking on Intel 8088 and 8086 processors? That work continued and eventually hit the market under names like like Concurrent CP/M, Concurrent DOS, and Flex OS. And not only did they multitask, they included varying degrees of MS-DOS compatibility. The compatibility was not 100%. Simple text mode applications would run, but more complex text mode applications like a word processor did not.

Ironically, when IBM first released its PS/2 line with a new advanced bus architecture and faster processors, the only operating system that took full advantage of those capabilities when it was first released came from Digital Research. Flex OS did everything people expected to be able to do with OS/2 on day one. But everyone ran DOS 3.3 anyway.

It was time for a different approach. A more direct approach, including a new name to go head to head with Microsoft.

Enter DR DOS 3.31 and 3.4: The 1988 Comeback Special

After 1987, as OS/2 floundered and no one seemed to care, and MS-DOS stagnated, Digital Research decided to take a different approach. Maybe FlexOS was too ambitious and the market wasn’t ready for that yet.

So Digital Research gave the market what it wanted. A dumbed down, more overt MS-DOS clone.

That’s what DR DOS 3.31, released in 1988, attempted to be. The version number suggested feature parity with MS-DOS 3.31. It delivered everything people bought (or pirated) MS-DOS 3.31 to get. But it also included enhancements like better error messages, command line history along the lines of what you get with DOSKEY in later versions of MS-DOS, and even some security features like the ability to password protect files and directories.

The approach worked. Something less ambitious but more compatible with MS-DOS/PC DOS 3.x and directly addressing the most obvious shortcomings of DOS put Digital Research back in the game. They even scored some OEM deals to bundle DR DOS with new computers in place of MS-DOS.

Version 3.4 quickly followed, with some minor enhancements, but the major difference was marketing. The version number suggested it was an improvement over MS-DOS 3.31, which it was. And Digital Research sold it at retail. That was new. IBM sold PC DOS through its authorized dealers, but by 1988, there was a whole other world of software retailers beyond IBM’s dealer network. Microsoft didn’t sell MS-DOS at retail at all.

DR DOS Version 5

DR DOS

DR DOS version 5 gained a cult following among power users and forced Microsoft to up its game and release MS-DOS 5.0.

Since MS-DOS and PC DOS version 4 was a bug-ridden flop, Digital Research skipped that version number altogether. In 1990, Digital Research released DR DOS 5.0. It ended up being very similar to MS-DOS 5.0, which came out a year later. That’s no coincidence.

Feature wise, this release was very similar to the Microsoft product. But it reported itself as DOS version 3.31 for compatibility purposes. So you can sometimes encounter software that will not run on DR DOS 5.0 and it will give the confusing error message that it requires DOS 5.0 or later in order to run.

But aside from that, it was a big step up from the DOS 3 family. It delivered memory management, a disk cache, a graphical file management shell, and a file transfer utility. It even offered memory management on 286 systems, so it had some capability out of the box that even MS-DOS 5.0 didn’t have.

Digital Research ran a campaign called “Toss your DOS,” offering a rebate if you sent in a copy of Microsoft or IBM DOS. Critics gave it good reviews and power users took a liking to it.

DR DOS Version 6

After Microsoft released DOS 5.0, Digital Research followed 3 months later with its own version 6, which included disk compression and undelete functionality, and even task switching. It wasn’t multitasking, only switching. But some of that lost progress from the early 80s was coming back.

Since Digital Research was getting harder to compete with, Microsoft decided to try some FUD. Microsoft inserted an error message in the beta version of Windows 3.1 if it detected DR DOS. It took 6 weeks for Digital Research to develop and release a patch to circumvent the code, and Microsoft disabled the code in the released version of Windows 3.1, but to some extent, the damage was done.

DR DOS always had some compatibility issues with certain games, though it generally worked well for business software. Arguably it ran Windows better than Microsoft’s own product did, but with Windows gaining importance, word of that error message scared off some would be purchasers.

The Novell buyout

Networking giant Novell purchased Digital Research in late 1991, and continued to sell DR DOS 6 under the same name. They just changed the packaging to include the Novell logo and branding.

Two years later, they released Novell DOS 7, an enhanced DR DOS. It was more ambitious in some ways. Unlike previous versions, which were basically a DOS 3.31 clone with enhanced utilities included, Novell DOS 7 was internally much more similar to MS-DOS 5 and MS-DOS 6, while continuing the tradition of bundling better utilities.

It was a good product, but it sold poorly, at least in the United States. It was more popular in Germany.

Novell had grand designs of competing directly with Microsoft all across its product line. To that end, they bought Wordperfect and other established products. But ended up losing on every front, including networking, where it had the head start.

Caldera Open DOS and DR-DOS

Novell sold its DOS product to Linux vendor Caldera in 1996. Caldera was one of the leading Linux distributors at the time, and wanted to bundle better DOS compatibility than competitors like Red Hat would be able to deliver.

Caldera released the source code with a limited license. Critics said the product was neither open nor free. You could download it without paying for it, but you were restricted in what you were allowed to do with it. The initial product was called Caldera Open DOS. But soon Caldera reverted back to the older, more familiar name.

The Microsoft lawsuit

Some people thought it was a front. In the February 4th, 1997 issue of PC Magazine, John C Dvorak wrote some speculation about the product, including the following:

I know one writer who claims that he has an original version of MS-DOS, version 1.0, which was created by Seattle Computer Products and bought by Microsoft for $50,000. The writer claims that if you hit a specific series of keys, you’ll get Gary Kildall’s name and some other indications that some of the code was actually lifted from CP/M. Apparently this was fixed rather quickly and certain secret deals were made between Digital Research, IBM, and Microsoft regarding CP/M, CP/M-86, and the product that eventually became MS-DOS.

I had the opportunity once to ask Jerry Pournelle if he was the writer Dvorak was talking about. He confirmed it but would not elaborate, only saying, “I know that story.”

Pournelle did speak about it on a podcast a number of years later, but said he could not remember the specific incriminating keystrokes.

As a writer, I want to believe the story, and efforts to disprove the existence of that keystroke have their problems, but then again, those early versions of DOS have been thoroughly disassembled and studied at this point.

Pournelle would have been in his late 40s when he observed this, and in his 60s when he told the story. Is it possible that he conflated the Easter egg Microsoft put in Basic when you typed the command wait 6502, 1 with MS-DOS and misremembered?

The $280 million settlement

At any rate, Caldera did indeed sue Microsoft, but it was for bundling DOS into Windows 95 rather than selling the two separately. And they did produce a version of DR DOS that was capable of running Windows 95 in place of the DOS Microsoft included and did its best to keep hidden behind the surface.

Microsoft and Caldera settled in January 2000 for an undisclosed sum, which was eventually revealed in November 2009 to be 280 million dollars.

Caldera took on a life of its own, splitting its Linux business and its DOS business. The Linux business eventually merged with UNIX vendor SCO, and even took on that name. SCO became infamous for suing IBM over its contributions to the Linux kernel, in spite of its history as a Linux vendor itself.

Caldera Embedded Systems and Lineo

DR DOS went through some minor revisions after it spun off from Caldera’s Linux division. It saw some success as an embedded operating system, but DOS of any sort was a tough sell in the late 1990s. IBM was still selling PCOS at that time, and not with great success either.

They licensed it for distribution with other products that needed to be able to distribute media that booted into a functioning DOS environment, such as hard drive utilities. And it worked well for this because it had been updated to include Fat32 support. But its days as a consumer operating system were behind it.

Devicelogics

When Lineo was bought out in 2002, some of its former managers bought the rights to DR DOS and continued improving and marketing it. But DR DOS faded after 2005, when it was discovered to be using some components, uncredited, from the FreeDOS project. Its last version was released in July 2011.

Legacy

DR DOS did indeed exact some revenge in the form of that $280 million settlement, but it did not result in tying up that loose end about Gary Kildall’s secret Easter egg.

And in many ways, for a while, it was a better DOS than Microsoft’s. It forced DOS to improve after years of mediocrity, and to a degree it lit a fire under Windows and OS/2 as well. That’s the beauty of competition.