The Rise of Microsoft Windows Part 1: Windows 1.0

The long, bumpy road to the first release of Windows

Today we will be looking at the origins of Windows, how it started as a directionless project at Microsoft, its struggle just to get released, and how it finally staggered onto the market, only to face a very uncertain future. 

This is the lightly edited script for the video of the same name, over on our YouTube channel.

1981

So let’s hop back to the computing world as it was in 1981. At this time the IBM PC had just been released and its MS-DOS operating system, for all of its many weaknesses and deficiencies, was rapidly rising towards the industry prominence it would enjoy throughout the 1980s. However, MS-DOS’s measly 27 system calls, 640k RAM barrier, cryptic commands, and lack of multitasking were pain points for not only its users, but also the software companies that were springing up to supply applications to the rapidly expanding PC market.

MS-DOS did so little that any application written to run under it had to handle a number of things on its own, such as printing, that modern operating systems take care of. Forced to continually reinvent the wheel, application developers were just as frustrated as the end users. For their part, the end users struggled to master the complexity of MS-DOS’s command line only interface, with its extremely unintuitive commands.

What many felt would be a desirable step forward was some form of graphical user interface, or GUI for short, something with windows that hid a lot of the complexity of MS-DOS from the user, and had a way to interact with it that was superior to memorizing long strings of text commands. This did not necessarily mean a mouse by the way, and there were many that felt that the command line was the way to go, it just needed more polishing.

By 1981 Xerox PARC’s graphical user interface was well known throughout the computer science industry, and was quite mature, having been around since the Alto’s creation in 1973. The Alto however was not a commercial product, it was created solely as an in-house computer for PARC, and while around 500 wound up in various universities, the other roughly 1500 or so stayed with Xerox and its subsidiaries. 

The Alto’s influence was significant, as it had brought to practical usage many of the concepts that the pioneering computer researcher Douglas Engelbart had spoken about and demonstrated in his 1968 landmark presentation on human/computer interaction, usually referred to as The Mother of All Demos. But since the Alto was a purpose built, expensive, research machine that was filled with custom hardware, it was not a product designed to bring GUIs to the masses.

However 1981 would be an important year in the personal computer industry, with the start of two projects designed to bring a graphical user operating system to existing IBM personal computers, no custom hardware needed. Additionally Xerox would finally make a serious attempt to commercialize the Alto’s technology.

And, although the projects had been ongoing for several years, it's worth mentioning that Apple Computer was also working on two major projects in 1981, both attempting to bring a full GUI to a personal computer that was built from the ground up from both a hardware and a software perspective to handle it. There was a lot of buzz around GUIs...but nobody was exactly sure how it should be implemented, or even if a GUI was anything more than a shiny distraction that added little to a computer’s usability.

Apple’s first new computer that was being developed was designed from the ground-up to use a graphical user interface. I’m talking of course about the revolutionary Lisa, not its more famous relative, the Macintosh, which was still a project very much in flux in 1981 and still had lots of...shall we say drama ahead of it to go through. The Lisa project had been started in 1979 and its graphical user interface had gone through several iterations, including one that didn’t use the mouse at all.

The famous trip to Xerox PARC which occurred shortly after the Lisa project began inarguably led to a shift in the Lisa’s development, although just how much is a matter of some debate. The debate ranges from those who say that all the Lisa did was clone Xerox’s GUI, adding little originality, to others who say that the trip had far less effect than is usually claimed.¹

But the Lisa would not be released until 1983 and the Macintosh wouldn’t make it to market until 1984, and neither would find success in the marketplace. The Lisa flopped so hard that leftover models were eventually reworked to run the Macintosh operating system and the Macintosh itself, although influential, sold poorly for years.

Xerox actually launched the very first personal computer to come with a full graphical user interface in 1981, called the Xerox Star or more prosaically, the Xerox 8010 Information System. It was remarkably advanced, boasting a rich set of features that in many ways were well ahead of its time. However, as was sadly typical of Xerox’s muddled attempts to market the fruits of PARC’s efforts, the Star face planted immediately upon release.

Its problem was that it was not really envisioned as a standalone system, rather it was supposed to be more of a turn key system for an entire office, with multiple Stars connected together via fast Ethernet, sharing files, printers, and freely emailing amongst each other in a way very comparable to something like today’s Slack or Teams environments. While technically a personal computer meant to be used by a single individual, the price of all of this technology was a painful 75,000 dollars for a base setup of Stars with a single laser printer to share, plus 16,000 dollars for each additional Star.

This wasn’t exactly what the market was looking for and the Star quickly sank into obscurity, although its advanced operating system would see a sort of half life with the eventual release of Xerox GlobalView in the late 80s, a Star emulator designed to allow software created for the Alto and its successors to run on Windows, Solaris, and OS/2 software. GlobalView initially required a hardware expansion card, but the final release, GlobalView 2.1, in 1996, was able to run solely under software emulation. Ultimately, the Star left very little impact, being an expensive dead end. 

What the market was looking for was a graphical user interface of some sort that didn’t require expensive custom hardware, worked with the rapidly expanding base of DOS programs, meaning that it had to coexist with DOS somehow, and also added significant functionality to IBM PCs and by extension the nascent IBM compatible market, while also being comparatively easy to develop for.

Additionally, given the vast amount of original IBM PCs with 8088 processors and far less than 640k of memory, a graphical user interface that could run on these lower end machines would have a far larger market than one that required a 286 or 386 processor. This was an incredibly difficult problem to solve and over the next few years multiple companies’ attempts to meet these requirements were to fail miserably. 

VisiCorp, the publishers of VisiCalc, the world’s first spreadsheet for the personal computer, which we covered in an episode of Retro Tech Bytes, had a plan of their own. In 1981 they started development of a graphical user interface called Visi On that was designed to run on IBM PCs on top of DOS. Visi On came to market two years later in 1983 and was quite advanced, boasting overlapping windows,  a consistent interface for applications running under it, a form of multitasking for DOS programs, was fully mouse driven, and applications could do basic data sharing such as cut/copy/paste between each other.

Unfortunately Visi On was doomed by a number of inherent problems. First of all its system requirements were extremely high, requiring a minimum of 512k of RAM and a hard drive. 512k of RAM was a massive amount for a 1983 PC, and hard drives were still extremely expensive and rather uncommon. As a matter of fact, IBM didn’t even offer a PC with a hard drive until March of 1983 anyhow, with the release of the XT.

Visi On was also expensive to buy, costing 495 dollars at release, plus a 250 dollar mouse. Furthermore, Visi On was difficult to develop for as it was not possible to write Visi On specific applications under Visi On itself. Rather, the programmer was required to use a Unix development environment, something that would run on an expensive workstation that cost around 20,000 dollars,² and then transfer the compiled program to an IBM PC running Visi On.

This certainly wasn’t unheard of in that era, and given how slow personal computers of that era were, it was actually fairly common to do software development on far more powerful workstation computers. However, software developers could always choose to do development straight on a PC, Apple II, TRS-80, C64 etc and just accept the much slower environments and more limited development tools in exchange for a drastically reduced cost of entry.

With Visi On, you either shelled out for an expensive workstation...or you didn’t develop for it at all. Unsurprisingly, Visi On barely staggered along for about two years, going defunct in November of 1985 when VisiCorp merged with Paladin Software and most software products were cut.

Then there was the second graphical user interface started in 1981, a windowing shell for MS-DOS called Interface Manager that Microsoft had started developing in September of 1981³ without much of a real direction, apart from a few bullet points from Bill Gates. As a side note, there is an interesting photo that ran in the Seattle Times on February 14, 1982. The photo is of Bill Gates and Paul Allen at Microsoft, and very interestingly you can just make out the words “Window manager” on the blackboard behind them.⁴ 

Interface Manager had somewhat grown out of two early programs that Microsoft had created, a VisiCalc competitor called MultiPlan and a word processor called Multi-Tool Word. Each of these programs attempted a sort of proto-GUI interface with descriptive menus, or at least descriptive menus by the very loose standards of the early 1980s. In fact, Multi-Tool Word actually used the very first Microsoft mouse, which retailed separately for 195 dollars, although the early releases did come with one bundled. Multi-Tool Word was an application, not an operating system or GUI shell, but Microsoft implemented windows in it, allowing for up to eight documents to be opened simultaneously.⁵

Multi-Tool Word also had a form of WYSIWYG, or “what you see is what you get”, with a mode that tried to display the document exactly how it would look when printed, complete with fonts. As it started its slow road to eventually dominating the word processing market, Multi-Tool Word was eventually renamed to the much less clunky Microsoft Word.

As mentioned, Bill Gates had initially specified only a few requirements for Interface Manager, key ones being that Interface Manager needed to have a rich set of drivers that would keep applications separate from hardware, have a far more robust toolkit for programmers to take advantage of than MS-DOS’s puny 27 function calls, work in graphics mode instead of trying to fake it in text only mode, support WYSIWYG, and push applications towards a standard appearance.⁶

As originally envisioned, Interface Manager was not actually an operating system, but rather a graphical shell for DOS that would simplify the job of managing drivers and making a consistent interface across applications.

As the legendary John Dvorak put it in a 2020 retrospective, “The thing wasn't even designed as an OS but a shell program that would have a zippy GUI and be able to manage all the device drivers through a common API. People forget that during the 1970’s and 1980's device drivers for peripherals were a nightmare. Most of them the user had to code themselves. “Interface Manager”...would smooth out the problem.”

Basically Interface Manager would handle communicating with various peripherals such as printers, freeing application developers from having to spend enormous amounts of time struggling to write drivers and frequently shipping multiple floppy disks of drivers with their programs, just so people buying the latest version of WordStar could have the one driver they needed for their printer. But again, as initially envisioned, Interface Manager would not be a full operating system.

As a top tier programmer himself, Gates knew the value of making software development easier and more full featured for developers as that would help drive Interface Manager’s adoption, and so he wanted Interface Manager to provide developers with a good experience. Get more applications that people wanted to use, and more users would flock to Interface Manager, making a larger pool of customers that application developers would be incentivized to develop for and so on and so on in a virtuous cycle. 

Originally, Interface Manager had been planned to base its appearance off of Multiplan, with alphabetical lists of commands at the bottom of the screen.⁷ However in 1982 the decision was made to move to something close to what the Xerox Star, Macintosh, and Lisa did and use pull down menus and dialog boxes.

This may or may not have been directly related to Bill Gates’ insider knowledge about what was going on with the Lisa and Macintosh projects, since he had a lot of access to those projects given Microsoft’s status as a key developer for Apple. Gates had also been given a personal look at a very early Macintosh prototype⁸ during the summer of 1981 by Steve Jobs himself, so it's very possible that the tour kindled his interest in starting a GUI for IBM computers. 

However it's also equally possible that Gates’ was simply taking a page from the Xerox Star itself or was tapping into his own pool of former Xerox PARC employees that Microsoft had hired. Gates had started his “PARC poaching” by luring away legendary programmer Charles Simonyi from PARC in 1981 who became enormously influential to Microsoft in the 1980s and was a key asset in the development of Multi-Tool Word and Interface Manager, where his PARC expertise in graphical user interfaces was critical to developing their look and feel.

Gates also eventually tapped another former PARC member, Scott MacGregor to oversee Interface Manager’s development and in 1983 he hired yet another PARC alumni, Leo Nikora to be Interface Manager’s product manager. 

Basically Gates had enough PARC people in-house that he probably didn’t need the tour Steve Jobs was given in order to know quite a bit about PARC’s innovations in user interfaces. Additionally, Gates had seen to it that Microsoft bought one of the first commercially available Xerox Stars, spending around 100,000 dollars for a setup that included a laser printer.

My source for this,  Hard Drive: Bill Gates and the Making of the Microsoft Empire says Microsoft only bought one Star,⁹ but a Star and a laser printer combined were nowhere near 100k so my suspicion is that Microsoft bought a full setup of at least five Stars at around 16k a pop, plus a laser printer for them to share.

Whatever the number purchased, the Stars were primarily there for research, or as former Microsoft programmer Bob Wallace said, “The Star wasn’t a real popular machine, but it was there to play with if you wanted to learn something about the user interface principles.”¹⁰ 

So in addition to having multiple PARC experts on staff, Microsoft also had Xerox Stars to study, again meaning that Gates definitely had enough knowledge and resources in-house to know pretty much all of the ins and outs of Xerox’s graphical user interface. 

1982

At Comdex in Fall of 1982, VisiCorp gave the first public demonstration of Visi On, and Gates’ immediately sensed a possible significant danger to Microsoft’s plans. He sat through three complete Visi On demonstrations, even flying Charly Simonyi down from Seattle specifically to see one of them.¹¹

Worried that Visi On would grab control of the GUI market for IBM computers, Gates made the decision to ramp up development on Interface Manager and then as soon as possible, start hyping it as being far closer to completion than it actually was, and with far more modest system requirements that it would turn out needing.

At a computer conference in January of 1983, Gates started dropping vague hints that Microsoft was working on a graphical user interface for PCs that would be better than, and also ship before, Visi On. It wasn’t until weeks after this announcement that a prototype of Interface Manager was even able to run on a PC at all and had mouse support added.¹²

Given that Microsoft had been working on it for several years at this point, presumably development had so far solely been occurring on a minicomputer development system, perhaps running an 8088 emulator. 

1983

April of 1983 saw the first public demo of Interface Manager, a demo that was such a mirage that it literally became internally known at Microsoft as the “smoke-and-mirrors” demo.¹³ Created by two programmers, Rao Remala and Dan McCabe the Interface Manager demo consisted “of a screenful of overlapping windows apparently running simultaneous programs, none of which actually did anything. Remala had cobbled the thing together by borrowing from graphics routines Microsoft had done for its Macintosh Chart application.”¹⁴

Remala had created the window manager for the demo, while McCabe had handled the graphics.¹⁵ The whole show was designed to give the illusion that Interface Manager was way closer to a release than it actually was...because at this point there was precious little more to Interface Manager than the demo. But the demo gave the industry the perception that Interface Manager would be released before too much longer, which was a double edged sword as it set expectations rather unhealthily high.

This demo would live on in Microsoft, as it was trotted out whenever one of the clone computer makers Microsoft was courting expressed any skepticism about the possibility of a slow 8088 processor handling a full windowing, multitasking operating system.¹⁶ By this point it was clear that Interface Manager needed to be a full operating system in order to leapfrog the competition, although I’m not actually sure exactly when that decision was made, as no resource I consulted mentioned it.

Microsoft pushed Interface Manager hard to the clone manufacturers, who were all already licensing MS-DOS from Microsoft. Microsoft account managers and sales reps pushed Interface Manager aggressively. The tactics Microsoft used to try to leverage its control of MS-DOS into sales of Interface Manager were extremely...shall we say used car salesman-like.

The excellent book Gates describes their tactics as follows: ”If you like DOS you’re gonna love Interface Manager. If you like DOS, this is the only 100 percent guaranteed compatible windowing program. If you like DOS and BASIC, we’ll turn the deal into a hat trick that gives you Interface Manager. Oh and you’ll probably need a mouse. We’ve got that too. It's easy: Sign here.” Almost every bullet point feature of Interface Manager was specifically designed to undercut Visi On as well.

Visi On required a hard drive, whereas Interface Manager could allegedly run on the two standard floppy drives every IBM compatible came with. Visi On required a lot of RAM, Interface Manager would supposedly only need 192k. Visi On was expensive, Microsoft would throw Interface Manager in for basically free with a lot of clone MS-DOS deals.

According to one source, the biggest licensors of Interface Manager got it for 8 dollars a system, or 24 dollars a copy. If a smaller manufacturer committed to a 20,000 dollar commitment, they paid 90 dollars a copy, which was still a lot cheaper than Visi On’s pricing.¹⁷ 

 It was cutthroat marketing at its finest, and was technically illegal under antitrust law, which forbids tying the purchase of one dominant product to another one. However Microsoft somehow managed to stay juuuust on the right side of not explicitly saying cloners needed to license Interface Manager in order to get MS-DOS and somehow got away with it.

However all of these manufacturers who were basically strong armed into signing licensing deals for Interface Manager were certainly going to expect it to be delivered on time and were not going to be happy if it wasn’t delivered when Gates promised it would be. And the chances of Interface Manager being delivered on time were…shall we say unlikely.

In August of 1983, with the pressure to get Interface Manager out the door steadily ramping up, Gates personally interviewed and poached away from Xerox 26 year old Scott MacGregor. MacGregor was a good friend of Charles Simonyi, who had personally recommended him to Gates as a good head for the Interface Manager group, which was finally given the name of the Interactive Systems Group on MacGregor’s first day of work.

MacGregor’s formal title was chief technical architect for Interface Manager. MacGregor was not impressed with Interface Manager when he saw it, nor was he impressed with Bill Gates’ understanding of what was involved with creating a full GUI. “He had no idea of what a windowing system was. He thought of it as a collection of subroutines that applications would incorporate to do windowing stuff, as opposed to a windowing environment.”¹⁸

Still, MacGregor had a job to do, and he threw himself into getting Interface Manager’s development moving along, although he knew that there was no way it would be able to be released the following year as Microsoft and Bill Gates were promising. McGregor also loftily compared himself to Prometheus, bringing PARC’s fire to Microsoft’s mere mortals.¹⁹

Before Microsoft’s GUI could be formally announced, it needed a new name as Interface Manager was not deemed suitable. Several other names were considered, among them Microsoft Desktop, which was the name Jeff Raikes, Microsoft’s director of applications marketing, favored. However the name that won out was a simple reference to the key component of a graphical user interface, a name suggested by Rowland Hanson, who had joined Microsoft as vice-president of corporate communications in early 1983.

Hanson was in charge of all advertising, packaging, merchandising, and anything else that had to do with marketing Microsoft products and he felt that every Microsoft product should have “Microsoft” as part of its official name, and all marketing should use it to help get Microsoft as a brand cemented into people’s minds. As Hanson says “It appeared that there were going to be multiple systems like this on the market. Well, we wanted to have our name basically define the generic”²⁰ for windowing based graphical user interfaces.

So he gathered his team and they started looking at all of the articles that had been written about GUIs that were either on the market already, or were close enough to have had a lot of information about them revealed. In Hanson’s words, “I told my team to go out and gather all the editorial that’s been written about all these other GUIs that are out there: “We’re going to look at all this stuff, and we are going to look for commonality in how these GUIs are being talked about.” So we taped everything to the wall - hundreds of articles around the room - and we started to find a very consistent theme between all these GUIs: they were not being described as ‘graphical user interfaces’; they were being described as ‘windowing systems’. And I said, “If these are all windowing systems, there’s only one name we can pick: we have got to be Windows.”²¹

Hanson got his way and thus Microsoft Interface Manager became Microsoft Windows. This change was made so last minute that there was barely enough time to update all of the PR materials. Meanwhile VisiCorp began shipping Visi On in October of 1983 and another company, Quarterdeck, announced that it too was working on a graphical user interface for MS-DOS computers.

In light of Gates’ proud announcement at the beginning of 1983 that Microsoft would be the first to release a GUI for IBM computers and compatibles, Microsoft was was now looking like its GUI would merely be the third entry in a crowded market.

Desperate to try to get some good publicity, on November 10, 1983, at the Helmsley Palace Hotel in New York City, Gates publicly announced that Microsoft Windows would be released in early 1984. Gates also predicted that by the end of 1984, over 90 percent of MS-DOS computers would be running Windows. Even though Windows was far, far away from being usable, Gates was concerned that waiting longer would see customers flocking to a different GUI, such as Visi On.

As Hard Drive: Bill Gates and the Making of the Microsoft Empire puts it: “He had learned that one way to prevent potential customers from flocking to a competitor’s product was to announce that your company was working on something even better. It was a tried-and-true IBM gambit that worked well when customers looked to your firm to set standards; they usually would gladly wait for the market leader’s product to come out.”²²

This industry practice was so prevalent that it's where the term “vaporware” came from, coined by magazine Infoworld. Another Microsoft manager later said “There seemed to be this notion that since all of our competitors were announcing products that were vaporware, we had to have one too.”

A Microsoft press release from the November rollout also firmly stated that Microsoft would not compete with its licensees, saying “Microsoft Windows is an OEM product, and each OEM is expected to set its own pricing, Microsoft Windows will be sold like MS-DOS, adding little or no cost to the system. We have no plans [to sell a retail version] at this time.”²³

Gates also, being privy to the still secret Macintosh project, firmly believed that it was going to shake up the industry when it was released, especially if it could keep its initial target price of under 2000 dollars. The Lisa had caused a major stir when it was released, but had sold poorly due to its hugely expensive asking price of 10,000 dollars, however the Macintosh at Jobs’ original target price of two thousand dollars could really bring the concept of a graphical user interface into the mainstream of consumer sentiment and demand.

If Gates could give those consumers a GUI that ran on their existing IBM compatible machines, the Macintosh could indirectly drive massive sales of Windows. Additionally, announcing Windows in such a public fashion prior to the Macintosh announcement would help drive the perception that Microsoft was, as Hard Drive puts it “surfing the curl of the breaking GUI groundswell, not languishing in the backwash.”²⁴ 

The Windows announcement was, according to Hard Drive, “the most elaborate product introduction ever witnessed in the industry”²⁵ up until that point. It lasted most of the morning, and featured multiple Microsoft employees putting a prototype of Windows through a very detailed demonstration.

The real standout wasn’t the GUI itself as nothing Microsoft was showing was revolutionary or new, what was surprising was the fact that Microsoft was claiming that Windows could do all of this on a normal basic IBM PC. However once again the extensive Windows demos being shown represented a very elaborate smoke and mirrors show.

In November of 1983, there existed a grand total of one actual Windows application, a very basic paint program similar to MacPaint. There were also a handful of applets, or very simple programs like a basic calculator and a bare bones text editor.²⁶ At this point, Windows was more of an elaborate tech demo than anything approaching a real product, and it was also still using the GUI from Multi-Tool Interface. 

Additionally, at this point Windows was based around block based graphics mixed with basic text, as opposed to the bit mapped graphics and proportional font based text of something like the Lisa.²⁷ Still, grandiose claims were made, especially to the developers that Microsoft wanted to entice to flood the market with Windows applications.

Product Manager Leo Nikora told the press that Windows would give developers a “development environment for MS-DOS so that applications can be graphically oriented” and promised that developers would be able to take advantage of “complete device independence and a graphical interface for MS-DOS.”²⁸

The advantage of developers being able to use an ordinary PC for development as opposed to Visi On’s requirement of an expensive DEC VAX minicomputer was also talked up as Microsoft pulled out all the stops to get as much hype as possible for a product that was primarily vapor at the moment, hopefully sucking the wind out of the competition.

It's important to point out that the Windows announcement also started to show the bad blood building between IBM and Microsoft as the cracks in their partnership began to widen. Microsoft was able to publicly talk up agreements with 24 major companies that promised to start including Windows as an option with their IBM compatible offerings, including such rapidly rising stars as Compaq, HP, Texas Instruments, Zenith, Tandy, Wang, and DEC.

These companies were very aware of the fact that IBM had tried to wipe out competition in the mainframe market before by abruptly changing its hardware or software standards. Based on the features, release date, and adoption rates that Gates was publicly proclaiming, these clone makers felt that Windows represented a great way to further escape IBM’s control of the PC ecosystem.

They already had hardware compatibility with the IBM standard, but were worried about IBM taking control of the operating system standard with their TopView GUI. If Windows took off and preempted TopView, then IBM would have lost control of both the hardware and the software standards for PCs, and at this point the clone makers were far more concerned about IBM’s power than they were about Microsoft’s.

Sure IBM could try to make a new standard, but nobody would follow them to it if Windows was firmly established. Or at least that seemed to be the thinking, and indeed that was broadly how it wound up working out some years down the road, as I covered in the Rise and Fall of the IBM PC series.  

Unsurprisingly, none of this sat well with Big Blue, who conspicuously distanced itself from Windows, and shockingly even signed a distribution agreement with VisiCorp for Visi On. This definitely came as a bit of a shock to Gates, who had assumed it would be daily straightforward to get IBM to start using Windows under similar terms to MSDOS.

As a matter of fact, the day after the Windows announcement Gates flew down to IBM’s PC headquarters in Boca Raton, Florida to pitch IBM on backing Windows for the second time. This trip ended in complete failure, with IBM vaguely mentioning an internally developed graphical interface that they were working on. This would wind up being the aforementioned TopView, IBM’s failed windowed, multitasking operating system that would die on arrival in 1985.

Just a couple weeks after Windows’ formal announcement in New York, Comdex 1983 kicked off in Las Vegas. And Windows absolutely everywhere, as Microsoft’s PR machine had spared no expense to put Windows front and center of everybody’s eyes. Taxis had Windows ad plates, attendees renting cars from Budget or Avis were handed Windows keychains to go with their rental, every hotel room had a bag slipped under the door packed full of PR material about Windows.

Windows was so prevalent everywhere in Las Vegas that according to one Microsoft executive, “You couldn’t take a leak in Vegas without seeing a Windows sticker that year.”²⁹ Pam Edstrom, Microsoft's head of public relations, even managed to get a bellboy to knock on the door of the VisiCorp suite and hand them one of the bags full of Windows materials.

According to her, “You could sort of hear”³⁰ the reaction emanating from the room. Given that Visi On would actually be released within a few weeks while Windows was still several years away, I’m fairly certain Visicorp had the next laugh, although the last laugh would actually be had by Gates once Visi On vanished from existence several years later, along with Visicorp itself.

Practical jokes aside, there wasn’t much to laugh about as Windows development was still mired in deep delays. And thanks to first the announcement, and then the Comdex hype, the pressure to deliver and the associated expectations, were rapidly increasing. It was also sometime in between the Windows announcement on November 10th, and the Comdex showing on November 29th that the decision was apparently made to make Windows a retail product, alongside the OEMs who had committed to shipping it preinstalled.

This decision was made due to IBM making it very clear that they would not be shipping Windows on any IBM machine. Without a retail version, Microsoft would be unable to tap into the massive amount of IBM PCs out there, especially the ones in businesses. One additional note of interest here is that the only real footage we have of Windows at this point comes courtesy of Dan Bricklin, of Visicalc fame, who went around Comdex with his camcorder and captured some of the earliest footage of Windows in existence.

As Windows struggled, the way too optimistic shipping date began to publicly slip and the industry began to take off the rose tinted glasses and start to reconsider their commitment. The stress of dealing with journalists and technology publications every time a delay had to be announced also fell quite heavily on Leo Nikora, Windows product manager.

As previously mentioned, Gates had hired Leo Nikora away from PARC in 1983. Nikora was a talented programmer, but the position of product manager was more of a marketing role. Nikora had no idea what he was getting himself into as he wasn’t in marketing at Xerox, but rather was an excellent engineer.

Nevertheless he would do his very best to learn marketing on the job as well as contribute to Windows’ overall development. In the words of Gates, Nikora was “a soft-spoken, gentle family man with dark, earnest eyes. These attributes supremely qualified him for his job at Microsoft, which was essentially to apologize for Windows.”³¹

Nikora also had helped manage the Xerox Star’s development and thus was an expert on GUI design, and had 17 years of programming experience. 

1984

 In February 1984, Microsoft held a conference aimed at software publishers and clone manufacturers. Attendees were expecting to get the technical information they needed to start developing applications and preparing computer systems to launch alongside Windows. Instead the 300 attendees, each of whom had paid 500 dollars to attend, were deeply disappointed to hear that Windows had received its first official delay, and was now said to be launching in May 1984. This left a bad taste in people’s mouths, and this bad taste got even more pronounced when May came and went without Windows, with Microsoft finally announcing that Windows would now launch towards the end of August 1984.

In light of the bad PR the delays were causing, Microsoft president Jon Shirley spoke to the press in July and attempted to put a positive spin on the struggles Microsoft was dealing with. Shirley was an extremely competent former Tandy VP who had been brought on by Gates in 1983 after Microsoft’s first president, James Towne, had been dismissed by Gates, who had decided Towne hadn’t been displaying enough enthusiasm for his job.

Shirley took a two fold approach with the press, first stressing that Windows was Microsoft’s primary focus, and thus was guaranteed to get done, and second pointing out that Microsoft had started shipping out development kits to programmers in May, as well as provided the clone manufacturers with the technical specifications needed for systems to properly run Windows.

This second point was designed to give the impression that Windows was very close to release, and versions of it were already out in the wild as it were, being used by developers to create exciting new Windows applications. This was...not terribly honest as Windows in mid 1984 was very much in flux still, and trying to nail down hardware specifications for it was basically like trying to nail jello to the wall.

The development kits that were shipped out would have been of very limited use given that Windows would change so much over the rest of development that an application that ran under the development version of it had a very high chance of not running correctly, or at all, under the final release. By this point, fewer and fewer developers were even interested in developing for Windows at all though so I guess you could say that this problem was affecting fewer developers by the day.

Meanwhile behind the scenes, Microsoft was doing some reorganizing. The Windows delays had exposed some serious structural problems in Microsoft’s development methodologies as it had taken a year of Windows development before everyone from Gates on down had really grasped the enormity of the Windows project and what a difficult task it was going to be to bring it to market.

The original projection for developing Windows expected a mere 6 man years total, spread across a handful of programmers, whereas by the time Windows 1 finally launched, it had taken the efforts of up 30 of Microsoft’s top programmers and 80 total man years.³² Towards the end of 1983 Microsoft had around 100 programmers so Windows was sucking up the time and energy of almost a third of the company's entire resources.³³

In the nascent software industry of the early 1980s this was a significant investment, even for a company as comparatively big as Microsoft and a miscalculation this significant showed just how disorganized Microsoft was in its projects. Matters were not helped by the fact that Gates oversaw every single development division, tended to launch new projects that never finished, moved programmers around from one project to another, and frequently would abruptly change specifications on various projects. This had all led to a culture of instability, and things needed to change or else Windows wasn't going to be the only major Microsoft project faltering. 

Additionally, feuds were breaking out between the Operating Systems Group, which was developing a new version of MS-DOS, and the Interactive Systems Group, which was handling Windows development. Since Windows was not originally conceived as an operating system, but rather, as one source puts it “a thin veneer over the operating system”,³⁴ some of the most advanced code needed for Windows was supposed to be done by the Operating System Group for the next release of MS-DOS and then given to the Windows developers for adaptation for Windows.

However the Operating Systems Group was falling behind its own schedule, a problem which was probably not helped by Windows vacuuming up so many resources, and thus MacGregor decided to have the Windows developers take on some of this coding in order to get it into Windows faster. This caused a major issue with the Operating Systems Group, whose head, Letwin, was furious. Matters were not helped by the fact that Microsoft's best programmers wanted to be on the Windows team, which for all of the stress, was viewed as the “exciting” place to be as opposed to working on boring old DOS.

This meant that MacGregor was frequently approached by programmers in other areas of Microsoft wanting to be moved to the Windows team, generating a lot of resentment from their management, especially Letwin. All of this squabbling was bad for morale, bad for productivity, and wasted a lot of energy that could have been spent more productively, yet Gates failed to intervene. 

There were also focus problems caused by Gates’ continual meddling, such as repeatedly changing his mind about whether Windows should allow windows to overlap, which is how the Liisa and Macintosh handled things, or force windows to tile without overlapping. Gates did not respect the chain of command, and injected himself into every decision he felt like, whether small or large.

Gates refused to allow the development schedule to include time for employees getting sick, unexpected bug fixing, and field testing, overriding MacGregor to delete such scheduled time and move up the release date accordingly. MacGregor would later describe this problem as stemming from Bill Gates' failure to understand “that when you have large and complicated projects with a lot of documentation and a lot of people, the trivial change takes a long time. You have to factor it through, change the code, change the documentation, and retest it, make sure the change works with all the other pieces of the program. There was this very real conflict between Bill’s desire to make Windows better, and the need to have a plan to execute, to actually get the product out.”³⁵ 

A manager who eventually left Microsoft described the chaotic atmosphere by saying that literally “everything was a crisis.”³⁶ Gates also eventually decided that since Windows was so far behind schedule, both MacGregor and Nikora should take on the additional burden of writing code, in addition to running their respective areas of Windows development.

This is not exactly considered a best practice for software development as, while nobody wants the proverbial clueless Pointy Haired Boss running things, managers of large software projects need to be able to focus on, you know, managing things. MacGregor and Nikora were both extremely talented software engineers, but stepping down into the trenches to crank out code would only further hamper the sprawling Windows project.

There was also a costly development change that occurred back when Windows was still known as Interface Manager. Although Gates had originally wanted Windows to be mouse only, with the keyboard not needed to control anything, after a year of development it was decided that Windows should actually support typing commands alongside using a mouse as there were concerns that consumers, used to using MS-DOS’s cryptic text based commands, would be reluctant to learn a new piece of hardware like a mouse.

Additionally, mouse only control would make the military reluctant to use Windows on any of its machines, or as MacGregor rather dryly put it, “There was concern about trying to use a mouse while driving a tank.”³⁷ In the end, the decision to add full command prompt support to Windows cost somewhere between three and six months of development time, which was just one of many painful setbacks to the always overly optimistic schedule.

June of 1984 saw a very interesting internal memo, jointly written by Gates and Ballmer that laid out Microsoft’s strategy for Windows, as well as explaining where the Macintosh fit into the plan.

The memo reads as follows: “Microsoft believes in the mouse and graphics as invaluable to the man-machine interface. We will bet on that belief by focusing new development on the two new environments with the mouse and graphics: Macintosh and Windows. This also makes sense from a marketing perspective. Our focus will be on the business user, a customer who can afford the extra hardware expense of a mouse and high-resolution screen, and who will pay premium prices for quality easy-to-use software. Microsoft will not invest significant development resources in new Apple II, MSX, CP/M-80, or character-based IBM PC applications. We will finish development and do a few enhancements to existing products. Over the foreseeable future, our plan is to implement products first for the Mac and then to port them to Windows. We are taking care in the design of the Windows user interface to make this as easy as possible.”³⁸

During August of 1984, as Microsoft struggled to decide how best to break the news that Windows would definitely not be launching anytime soon, a major internal reorganization was carried out. Things were reorganized into two main divisions, Operating Systems and Business Applications.

Steve Ballmer took over as head of the operating systems division and Ida Cole was recruited from her position as Apple’s marketing director to take charge of Microsoft’s applications division. Gates’ primary focus was now strictly long term product planning, defining software products, and determining Microsoft's overall development direction. The hope was that this restructuring would help Microsoft focus on its projects, especially Windows, and put an end to the turf wars.

Additionally, and somewhat quietly, this reorganization was meant to quell complaints from outside developers that Microsoft’s internal application programmers received an unfair advantage as they were not only privy to inside details of how best to program Microsoft’s own applications to work with Windows, but also received could easily see what competitors were doing with applications that ran under MS-DOS. 

Since any major change in an application generally resulted in a meeting with Microsoft to make sure that it would still run correctly under MS-DOS, Microsoft was widely perceived to have a massive advantage in knowing just what its competition was up to. Gates publicly proclaimed that this advantage would be nullified by the reorganization, claiming that there was now a barrier between the two divisions that would prevent Microsoft’s applications division from receiving insider knowledge from the operating system division and place everyone on a level playing field. The industry quickly dubbed this barrier Microsoft’s “Chinese Wall.”

If you believe that this wall actually existed in any appreciable way, I have some beachfront property to sell you in Arizona and I can get you a special two-for-one deal if you buy a certain New York City bridge as well.

Meanwhile back at Windows, Neil Konzen joined the development effort, specifically taking charge of designing the user interface. Konzen had been working on developing Microsoft applications for the as-yet-unreleased Macintosh for the past couple years and had very strong opinions about how Windows should look and feel.

He also felt strongly that Windows should make it easy to port applications over to the Macintosh, and also pushed to make Windows’ internals easier for programmers to work with, causing a significant amount of reworking of Windows code in a number of areas, resulting in friction building up between him and MacGregor. All of this pushed Windows’ development back even further, and in October of 1984, Microsoft was forced to announce that Windows would not be released that year, and in fact was being pushed back all the way to June of 1985. 

Leo Nikora once again had the unenviable task of justifying the delay to the press, a press which was getting increasingly critical of Windows interminable delays and was out for blood. Nikora's many years of programming experience and Xerox PARC background made him an excellent and reputable authority on the struggles Windows was facing, and he was undoubtedly a better choice to try to spin things for the press, versus using some PR person without a technical background. Unfortunately, even Nikora couldn’t do much to mitigate the increasingly bad press raining down on Microsoft’s head. 

Still, he valiantly tried to put the best face on things, explaining that Windows currently took up too much memory and required too much processor power to be viable for the 8088 processors powering the bulk of IBM machines and compatibles, and Microsoft felt they could not release it until they reduced its hardware requirements. His explanation was not greeted with much support from the press, with many publications roundly deriding Microsoft.

Some of the criticism went even further, with Journal American stating that Microsoft’s failure to deliver Windows was its first strategic mistake and the 1984 Christmas edition of Personal Computer stating that Windows was a solution in search of a problem, claiming that graphical user interfaces did nothing but cause unneeded complexity for end users, an opinion that was shared by fellow tech publication PC Product. Forbes also piled on, pointing out that VisiCorp had not seen any success with Visi On, and was on the verge of complete financial collapse and thus Microsoft Windows was very much a solution in search of a problem.  

But Nikora was telling the truth, at the November 1983 announcement event Windows had been talked up as only requiring 192k of RAM, and no hard drive, requirements that were specifically chosen to undercut Visi On’s far higher requirements. It had also been claimed that all DOS applications would run under Windows, with those not coded specifically for Windows, which would be the bulk of them, merely running in full screen mode and without access to Windows advanced features.

Microsoft said that Windows would basically shut down once what they called a “misbehaved” DOS program launched, then restart itself once the DOS program was exited. This was also a marked improvement over Visi On, but this was a lot of very specific, very difficult things to promise and the massive expectations and promises were forcing Windows ever further behind schedule as the team struggled to not deliver on all the lofty feature promises, but also get Windows working correctly on the weak hardware of an original IBM PC. 

Windows development had turned into a brutal slog as programmers wrestled to get the overdue, overhyped, and overburdened operating system out the door. Gates, never much for polite behavior, took out his stress on the Windows team more than any other team at Microsoft, or as Hard Drive puts it, “No Microsoft manager, man or woman, escaped the emotional tirades and verbal abuse of Bill Gates, but none took more of it than those working on Windows.” ³⁹

Gates had raised massive expectations with all of the hype for Windows, promising the world that Microsoft would deliver an extraordinary piece of software, something capable of bringing stodgy old IBM PCs into the shiny new world of graphical user interfaces, and the struggles going on to just get Windows into basic functionality, balanced against all of his promises, were enormously draining.

MacGregor later said “I don’t think Bill understood the magnitude of doing a project such as Windows. All of the projects Bill had ever worked on could be done in a week or weekend by one or two different people. That’s a very different kind of project than one which takes multiple people more than a year to do. There were 30 some odd people in the Windows group by the time we were fully staffed, which for Microsoft was the largest project they’d ever done.”⁴⁰ 

Additionally, the Windows development death march had started to majorly affect morale at Microsoft, with programmers no longer trying to get on the Windows team. Microsoft’s application developers were also resisting working on Windows applications. preferring to focus on developing Macintosh applications.

Somehow the Windows team kept up their overall belief in the importance of the Windows project, even though according to the book Barbarians Led By Bill Gates, “Almost all the key developers on the project thought of quitting or even tried to resign at least once. But not many actually did. Back then, the attrition rate was very low because people believed in the Windows vision. They believed that the software they were writing would revolutionize the world.”⁴¹ 

But by this point, they were about the only ones still believing in Windows, since outside Microsoft’s walls the Macintosh hype was building to enormous levels, leaving little room for Microsoft's oft delayed GUI.

Microsoft was still a major, if not the major, third party Macintosh developer and had been given considerable access to the Macintosh team, Macintosh prototypes, and developer resources. The buzz around the Macintosh just kept building, especially after the famous 1984 Super Bowl commercial, and Windows was just not generating the same level of excitement. Or really any excitement at all. Yet the Windows developers grimly kept moving forward.

There was a lot riding on Windows and its failure could take down Microsoft and leave them vulnerable to market disruption from a competitor. In a January 1984 article in Fortune magazine, the authors laid out the magnitude of Windows failing: “a lot is riding on Windows. If it fails to become an industry standard, Microsoft may not get another chance to take the consumer market by storm. Momentum is an ephemeral quality in any business, and in an industry evolving as fast as microcomputer software, it can be lost in the blink of an eye...Like other fast-growing companies racing to seize transient opportunities, Microsoft has devoted little time to develop the kind of management depth that will be needed to turn temporary victories into long-term dominance.”  The pressure was high, the stakes were higher, and Bill Gates’ ulcers were getting bigger. 

RAM was probably the biggest technical issue Windows faced, although getting multitasking working on a 4.77 mhz 8088 processor was definitely its own considerable challenge. DOS programs were used to having the entirety of a given PC’s RAM available to them.

Once a program, Lotus 1-2-3 for example, was launched, it would check the amount of installed memory and then grab all of it for its own use. At that point Lotus 1-2-3 was basically in full control of the entirety of the computer’s resources down to the hardware level, with MS-DOS relegated to a mere background role. What Windows needed to do was figure out a way to force a DOS application to only grab some of the available RAM, but make it think that it had all of it. 

With independent software developers now having been burned multiple times by Microsoft’s continually changing release dates for Windows, Gates himself wound up responding to at least some of them personally, taking at least some of the heat off of an increasingly burned Nikora who was getting very tired of making a circuit of software publishers over and over again, each time explaining that Windows would once again be pushed back.

In one letter to a software publisher from October of 1984, Gates said “Windows is the most strategic product that Microsoft is working on, we want it to be the environment of choice for the next generation of graphical applications. To achieve this goal, it must have the functionality and performance required by this new generation of graphical applications. This will not be easy; it will be a significant advancement of the state of the art.”⁴²

In this letter, Gates also tried a fairly clever bit of PR spin, telling the publisher that although the press was being told that Windows would be released in June of 1985, Microsoft actually would actually release it on April 15th.

This falsehood was Gates’ way of trying to spin an April release as actually coming in earlier than than its newest official release month. But given Microsoft's abysmal track record of success in meeting any Windows delivery date so far, the smart money was not exactly betting on the company to meet this latest deadline any more than it had the last few. 

Gates himself continued to be the cause of many of Windows’ delays, with Hard Drive stating “In the past, Gates’ strategy was always to get out first with a product and grab a share of the market and fix the problems later. With Windows, he inexplicably took just the opposite approach. He couldn’t bring himself to turn it loose until everything was perfect.”⁴³ 

Nikora himself clearly felt that this was accurate, saying in an interview: “Most of the reason Windows didn’t get out was because Bill kept adding functionality and changing the rules. I was saying ‘Look Bill, you’ve got to be out there, take what you’ve got and get it out there. You can always make it better, you can always improve it later, but get it out there.’ Bill’s feeling was ‘No, we’ve got to go out there with something that’s just head and shoulders above everybody else. It’s gotta be the right product.’”⁴⁴

The industry was also starting to attack Gates personally, not just Microsoft, questioning his business judgment and wondering if maybe he had lost his magic touch. 

Gates also kept ordering changes to Windows to make it more Mac-like, sometimes these changes required redoing already functional sections of Windows, and other times it required adding entirely new features. For example, Windows originally had proportional scroll bars, where the size of the scroll bar got bigger or smaller depending on how much of the document was visible.

Gates required that the scroll bar be redone to always stay the same size, matching what the Macintosh did. At one point Windows had full drag and drop enabled for copy/paste, using the mouse to implement this feature in a way that the Macintosh did not. Of course Gates wasn’t happy by this, and so this feature was completely removed. The book Barbarians Led By Bill Gates estimates that Gates’ attempts to force Windows to be more mac-like cost 7-8 months of development time, which based on everything else I read is probably a significant understatement.⁴⁵

Additionally, even by late 1984 Windows was still directionless in many areas, or as Barbarians puts it, “The developers still had not received a clear directive on whether Windows would subsume DOS, be an application that ran on top of DOS, or be just a little graphics subroutine library.”⁴⁶

Windows was a project without much of its own identity, and then Gates decided to throw another wrench into its development...TopView.

In brief, Gates was concerned about TopView supplanting Windows and thus issued a directive to add TopView compatibility into Windows, meaning that Windows could run on top of either MS-DOS or TopView. Fortunately for Windows, this requirement was dropped, but not before programmers had already wasted valuable time trying to hack it into Windows. 

In 1984 Gates reorganized Microsoft yet again, this time putting his most trusted lieutenant, Steve Balmer, in overall charge of the Windows project and piling the pressure on him to get Windows done or else. To quote the Digital Antiquarian, “Ballmer started burning through Windows project managers at a prodigious pace. The project acquired a reputation inside Microsoft as an assignment to be avoided at all costs, a place where promising careers went to die.”

In the end, Windows would burn through four product managers, with the fifth, Tandy Trower, only reluctantly transferring to the Windows division. Trower actually initially assumed that he was only being transferred to Windows because Gates wanted an excuse to fire him, saying “Nobody wanted to touch Windows. It was like the death project.” But there was even more bad news on the horizon. 

1985

As 1984 gave way to 1985, Microsoft had to switch up the entire plan for releasing and marketing Windows. The original reason for getting as many of the clone makers as possible on board was so that they could sell new PCs with Windows preinstalled as part of the package. However, no matter how hard they tried, Microsoft simply could not get Windows to run fast enough on the basic hardware that made up the majority of the PC market, which was heavily dominated by lower end, 8088 based machines.

Additionally all of the deals struck with clone OEMs were falling apart since Windows was so badly overdue that every company had pretty much lost interest. The original plan was for Windows to be OEM only. Then it had changed into a hybrid of OEM and retail. Now, reluctantly, Gates decided that Windows 1.0 would have to be released solely as a boxed retail product. Microsoft would just have to hope that enough consumers bought it for it to become a standard, pushing the as yet unknown quality of TopView to the side and getting the clone manufacturers back on board. 

Meanwhile, the Windows team was convulsing under yet another internal issue, as Steve Ballmer’s dislike of Nikora had flared into what was essentially open war. Ballmer was apparently in Nikora’s office practically every day engaging in shouting matches, pounding his fist on Nikora’s table, and calling him an “idiot” whose ideas were “complete jokes”.

Nikora, an excellent programmer and manager, who had arrived at Microsoft from a very successful career and was doing his very best as Windows’ project manager, eventually burned out and quit Microsoft, saying “I was burned out, I also left because Microsoft wasn’t quite ready for middle management. They thought they were ready, but they weren’t.”⁴⁷ 

And it wasn’t just Nikora who Ballmer was butting heads with, as Ballmer and MacGregor were also fighting. MacGregor took issue with how Ballmer, a non-programmer, attempted to manage the Windows programmers by continually ratcheting the pressure higher and higher on them, demanding they do jobs that should take six weeks in just half the time, and driving the group so hard that marriages started to get strained as spouses put extraordinarily long hours in at the Microsoft campus.

If you know your computer history you may see some parallels with how Steve Jobs, also a non-programmer, ran the Mac team, but Steve Jobs was able to make it work and Ballmer, although capable of being motivational, was not on Jobs’ level and lacked Jobs’ enormous charisma.

Hampered by his lack of a technical background, Ballmer struggled to communicate with developers, especially the super-stars that refused to respect somebody they viewed as having Bill Gates’ confrontational management style, without the technical skills and know-how to back it up.

And yet, MacGregor was himself struggling to get the Windows programmers to follow his vision for Windows. With his Xerox background, MacGregor wanted Windows to more closely replicate the Xerox Star, which had its origins in the 1973 Alto’s Smalltalk system. This system, for all its many innovations, was considerably behind the curve represented by the Lisa and Macintosh, both of which had taken the concepts of a graphical user interface much further than anything PARC had done.

One key design aspect that he refused to budge on was his insistence that Windows use tiled windows that could not overlap each other or be freely dragged around the screen like Macintosh or Lisa windows could. As Neil Konzen’s changes began to be implemented, bringing Windows closer to the Macintosh paradigm, MacGregor’s authority had waned. This, coupled with the constant fighting with Ballmer, plus the grueling development, led to MacGregor reaching his limit.

By spring of 1985, MacGregor had become so frustrated that he decided to leave Microsoft entirely. In his case, Gates himself interceded and attempted to convince him to stay, but there was no changing MacGregor’s mind. After MacGregor left, Ballmer moved into his empty office the following day.

Ballmer’s formally taking over MacGregor’s former position of chief technical architect caused awkwardness immediately with Windows fifth product manager, Tandy Trower, who had just moved in to his new position right as MacGregor exited the company. This put Trower in the weird position of having Ballmer as both his boss and subordinate, based on which hat Ballmer was wearing at the time. 

With Ballmer now running the Windows project, he was faced with yet another delay as it was now evident that Windows would not be on store shelves by June of 1985 as promised. Gates called Ballmer into his office one morning in the spring and started shouting at him, stating that Ballmer had better get Windows out before the end of the year or Ballmer himself would be out of a job. Ballmer took this challenge in stride, gathered the Windows team together, and said “Kids, we must ship this product before the snow falls.”

The final grueling sprint to get Windows out the door had begun, and things got a bit crazy as the youthful Microsoft programmers struggling with Windows gave up all thought of life outside of Microsoft’s walls. Working seven days a week and constantly exhorted by Ballmer, who had Gates breathing down his own neck, the Windows team threw themselves into their task with renewed determination, with a young Microsoft tester by the name of Gabe Newell even going so far as to spend an entire month in his office without leaving.

Newell worked around the clock and relied on napping in a sleeping bag whenever his exhausted body could no longer work without at least a modicum of rest. This went beyond even the standards of the other developers and Newell acquired the name “Madman” Newell from that point on. I wonder whatever happened to him…

Windows programmers also resorted to amping up their various ways of blowing off steam, typically using the down time while waiting for code to compile. And by blowing off steam I mean literally blowing stuff up, one night a mixture of sugar and saltpeter was used to make bombs and rockets, resulting in a police call, complete with bomb sniffing dogs. Nobody was caught, but it was a close call.

That didn’t stop a young programmer from attempting to make an even bigger bomb by pressing a cafeteria microwave into service to melt sugar so that the sugar/saltpeter mixture was even more explosive. Probably fortunately, the volatile mixture blew up in the microwave, reducing it to ruins, sending smoke through the building and bringing this particular way of stress relief to a halt.

The pressure was unrelenting, with one programmer later succinctly saying they “sweated blood trying to get this thing done” and another one saying “you felt like you were at the center of the universe. That was the motivation, that and just trying to get clean code out there. It was an invigorating feeling to be working for Microsoft. And all this pounding by Steve Ballmer, and yanking by Bill, was the price you paid to be there.”⁴⁸

Slowly, the disparate pieces of Windows finally started to come together. But it was a hard slog. The Macintosh and Lisa teams had been affording the luxury of a clean slate, with only one hardware configuration to target apiece. The Windows team was struggling to get Windows to run on a widely disparate amount of IBM compatibles, with more makes and models entering the market every week.

The struggle was to get all of the hardware that could theoretically be running MS-DOS, to be supported under Windows while also keeping MS-DOS from getting broken. This meant that in addition to writing Windows, the team had to also write drivers for printers, video cards, and as many hardware configurations as possible that needed to work under Windows just as well as MS-DOS.

When the spring Comdex came in May of 1985, Windows was still not ready. But it was feature complete, basically in beta, enabling what was termed a “Preview Edition” to be shown at Microsoft’s booth, and sent out to the many-times-burned outside developers that Microsoft was desperately hoping were still willing to write Windows applications. But Windows was still not ready for retail release. 

Finally, at the fall Comdex show six months later, in November of 1985, Windows 1.01 was shown, ready for purchase at last. At this point Windows was launching into a marketplace that viewed the graphical user interface with deep skepticism, IBM’s TopView was already a clear failure, Visi On was in the process of being buried as VisiCorp sold itself off the Paladin Software in November, the Lisa had already clearly failed, the innovative but hardware crippled and overly expensive Macintosh was struggling for traction, and Steve Jobs was already gone from Apple, licking his wounds and planning to start NeXT Computer.

Windows, the butt of so many jokes the past couple years, was not exactly launching into a favorable atmosphere. 

However a very smart and media savvy PR person at Microsoft by the name of Pam Edstrom had a clever idea to at least mitigate the jokes and ridicule by giving the industry a chance to make all the jokes they wanted...with Microsoft footing the bill. And so, as part of the launch of Windows at the November Comdex, Microsoft invited every single technology figure of any distinction to a “Microsoft Roast”, with legendary figure John Dvorak performing as master of ceremonies.

Gates and Ballmer cheerfully accepted a “Golden Vaporware” award from Infoworld’s Stewart Alsop, did a comedy routine together replete with some truly awful jokes, and finished up by dueting “The Impossible Dream” as the first 500 copies of Windows were rolled out onto the stage in a giant shopping cart, complete with dry ice billowing everywhere. At least there was supposed to be dry ice but given that the venue was in the extremely dry environment of Las Vegas what actually happened when dry ice was tossed into buckets of water was...absolutely nothing.⁴⁹

The evening was a success however, with lots of laughter from the attendees, and it seemed to soften perceptions toward Windows at least a bit. PC World wrote that “the future of integration [can] be perceived through Windows” and PC Week’s Jim Seymore wrote “I am a Windows fan, not because of what it is today but what it almost certainly will become. I think developers who don’t build Windows compatibility into new products and new releases of successful products are crazy. The secret of Windows in its present state is how much it offers program developers. They don’t have to write screen drivers [or] printer drivers; they can offer their customers a kind of two-bit concurrency and data exchange.”

However, Windows 1 was nowhere near usable enough to really catch people’s interest. It was great that Microsoft had finally gotten it out the door, but it was still a product very much in flux and seeking an identity. And it was dog slow on most systems, or in the more blunt words of one of the programmers who had been working on it for the past several years, “Windows was a pig.”

As Hard Drive puts it, “The program was too ambitious to run efficiently on most of the personal computers then in use. The machines lacked the memory and speed to take advantage of the best features the software had to offer. In addition, because of all the delays at Microsoft, there were few application programs available to run under Windows.”⁵⁰

Sure, Windows could technically run on an original IBM PC from 1981, provided it had at least 256k of RAM installed. However just because you could in theory do it...didn’t mean it was a good, usable, or enjoyable experience. For the end user on a typical PC towards the end of 1985, Windows added little and in fact was about as much fun to use as those marginally usable “Vista Capable” laptops that were foisted on quite a few people.

Windows 1 did have a lot of technical advancements going on under the hood, offering cooperative multitasking and virtual memory, and even enabling the usage of RAM beyond the 640k barrier...sort of. This last feature was done by basically treating RAM beyond 640k as virtual memory, same as a hard drive was handled. When applications were opened beyond the 640k limit on a system with more than 640k of RAM, let's say 1 Mb, Windows would swap data back and forth to keep whatever the active program was, loaded into the base 640k. 

This was not necessarily a smooth process, and unless you had a beefy 386, would quickly bring the computer to a standstill, or as PC Magazine put it, “the thrashing as Windows alternately loads one application and then the other brings the machine to its knees.” Windows 1 was technically an MS-DOS program, so programs that ran under it had to fit themselves alongside the available RAM left over from Windows itself.

These programs had their own file format for the executable, strictly for programs that ran under Windows though, if a normal MS-DOS executable was run Windows would simply launch the program and then basically stop running, only reopening itself once the MS-DOS program quit. It is fair to ask if Windows 1 was actually an operating system, or just a graphical front end for MS-DOS.

I tend to think of it as an operating system because it did have its own executables, as well as its own drivers that Windows applications could use to talk to hardware. However, since Windows required MS-DOS to be installed in order to function, and seemed to have relied on MS-DOS to handle file operations...you can certainly argue that it was more than a graphical user interface but less than an operating system.

Remember that all of the bundled deals with PC clone manufacturers that Gates had struggled to put together...had fallen through by this time. Windows 1 would be solely sold as a retail product. Microsoft priced in at an incredibly low 99 dollars in an attempt to get it moving and create some consumer demand that would spur developers to create Windows applications, but unfortunately it was an effort doomed to failure.

So few people bought it that essentially no consumer demand was created, meaning developers had no incentive to create programs for it. And since the vast majority of MS-DOS programs ran just fine under Windows, developers could keep making MS-DOS programs anyhow, secure that if one of the rare Windows 1 owners bought it, it would almost certainly run just fine for them. A similar compatibility issue would eventually wind up badly damaging OS/2 some years later, so Microsoft did eventually get to be on the other side of this issue..but that's a story for the OS/2 video.

One other thing of interest is the fact that the day after the roast, on November 22, 1985, Microsoft and Apple signed a critical, and confidential, agreement that would arguably prove to be Apple’s greatest misstep of the 1980s. In exchange for Microsoft’s promise to keep developing Excel and Word for the Macintosh, two programs that were desperately needed to fill out the Macintosh's slender shelf of available software, Microsoft gained the right to royalty free usage of graphical user interface technology developed by Apple for the Macintosh.

Part of the agreement put it pretty bluntly, “For purposes of resolving this dispute and in consideration of the license grant from Apple...Microsoft acknowledges that the visual displays in [Excel, Windows, Word, and Multiplan] are derivative works of the visual displays generated by Apple’s Lisa and Macintosh graphical user interface programs.”

This agreement was signed by Bill Gates and Apple’s new chairman, John Sculley. And it would have massive repercussions for Microsoft and Apple, just three years down the road when they faced each other in court in a battle that would have the entire industry’s attention.

And that wasn’t the only significant agreement to be signed in 1985, just a couple months prior to Windows launch, in August 1985, Microsoft and IBM had signed a Joint Development Agreement to start collaboratively working to develop MS-DOS’s replacement. This might seem a bit of a surprising development in light of the coolness that had risen in the relationship, but with the failure of TopView, IBM was unwilling to take on the job of replacing MS-DOS all by itself.

On Gates’ part, the struggle to get Windows out the door, the failure of every single agreement with a cloner to package Windows with their systems, and Windows’ own uncertain future as a retail product...well Gates wasn’t feeling terribly optimistic. Maybe it would be better to work with Big Blue and develop MS-DOS’s replacement together. The ramifications of the JDA and OS/2 would give Windows an extremely uncertain future as there seemed to be no point in continuing development of MS-DOS’s successor when IBM and Microsoft were working together on OS/2.

Windows had made it to its first release, but its future looked extremely doubtful.  

The story continues in The Rise of Windows Part 2: Windows 2x, available as a video or a post right here on the Substack.

 

1. https://www.folklore.org/StoryView.py?project=Macintosh&story=Busy_Being_Born.txt&sortOrder=Sort+by+Date&topic=Lisa

2. https://www.filfre.net/2018/06/doing-windows-part-2-from-interface-manager-to-windows/

3. Ichbiah, D., & Knepper, S. L. (1991). The Making of Microsoft: How Bill Gates and his team created the world’s most successful software company. Page 175

4. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 216

5. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 238

6. Ichbiah, D., & Knepper, S. L. (1991). The Making of Microsoft: How Bill Gates and his team created the world’s most successful software company. Page 176

7. Ibid

8. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 251

9. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 220

10. Ibid

11. Cringely, Robert, (1996) Accidental Empires, Page 219

12. Ichbiah, D., & Knepper, S. L. (1991). The Making of Microsoft: How Bill Gates and his team created the world’s most successful software company. Page 176

13. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 221

14. Ibid

15. Eller, M. (1998). Barbarians Led by Bill Gates: Microsoft From the Inside: How the world’s richest corporation wields its power. Page 23

16. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 225

17. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 241

18. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 240

19. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 254

20. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 252

21. Nicholson, M. (2014). When computing got personal: A History of the Desktop Computer. Page 124

22. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 257

23. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 241

24. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 257

25. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 258

26. https://www.filfre.net/2018/06/doing-windows-part-2-from-interface-manager-to-windows/

27. Nicholson, M. (2014). When computing got personal: A History of the Desktop Computer. Page 125

28. Ibid

29. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 242

30. Ibid

31. Manes, S., & Andrews, P. (1994). Gates: How Microsoft’s Mogul Reinvented an Industry--and Made Himself the Richest Man in America. Touchstone. Page 254

32. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 250

33. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 264

34. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 300

35. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 302

36. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 304

37. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 302

38. https://www.filfre.net/2018/06/doing-windows-part-2-from-interface-manager-to-windows/

39. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 294

40. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 295

41. Eller, M. (1998). Barbarians Led by Bill Gates: Microsoft From the Inside: How the world’s richest corporation wields its power. Page 51

42. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 306

43. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 307

44. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 308

45. Eller, M. (1998). Barbarians Led by Bill Gates: Microsoft From the Inside: How the world’s richest corporation wields its power. Page 57

46. Eller, M. (1998). Barbarians Led by Bill Gates: Microsoft From the Inside: How the world’s richest corporation wields its power. Page 61

47. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 309

48. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 312

49. Windows One. Happy Birthday.

50. Wallace, J. M., & Erickson, J. E. (1992). Hard drive: Bill Gates and the making of the Microsoft empire. Page 314