While researching various aspects of the history of computing, sometimes I come across what can best be described as “weird tales” — unsourced claims that sound interesting but are either provably wrong or there’s no evidence to support them. In the 1980s and 1990s, Bill Gates was a frequent purveyor of such weird tales, often designed to burnish the image of Microsoft.
Listening to Bill Gates, one might for example get the impression that he personally designed the IBM PC, even though Microsoft’s involvement with the actual hardware design was at best peripheral. At the same time, there’s often enough some kernel of truth in the implausible stories.
Some time ago, one of such weird tales caught my attention, the story of Microsoft putting the Wang word processing character set into the IBM PC. Here’s one version of the story:
They told us for this keyboard, they said Lexington had to use a certain layout. We wanted these function keys to be up here, that the Lexington layout forced them to be over here. We ended up with this funny big key over here because of Lexington. But here we’d put in our favorite fully-extended character set. Came up with some new ideas there. Put some special things in here that we thought we’d be able to use. In fact, we put in the Wang word processing character set because we weren’t sure whether to do a Wang-clone word processor, or start with our own approach at that time.
Bill Gates, Smithsonian Video History Interview
There is another variant of the story, similar but not the same:
Yeah, we were also fascinated by dedicated word processors from Wang, because we believed that general-purpose machines could do that just as well. That’s why, when it came time to design the keyboard for the IBM PC, we put the funny Wang character set into the machine–you know, smiley faces and boxes and triangles and stuff. We were thinking we’d like to do a clone of Wang word-processing software someday. Most personal computers back then didn’t even have upper- and lower-case characters.
Bill Gates, Fortune Magazine interview, October 2, 1995
First a side note about “Lexington”. No one appears to know of an IBM keyboard codenamed “Lexington”. On the other hand, it is well known that the original IBM PC keyboard layout came straight from System/23 Datamaster. It is also well known that IBM’s keyboards were manufactured in Lexington, Kentucky at the time (later Lexmark). Bill Gates most likely mixed things up, perhaps due to incomplete information he had.
Anyway… so smiley faces in the IBM PC character set supposedly came from Wang word processors. Yet Dr. Dave Bradley, one of the IBM PC development team members, says that he was present when the PC character set was sketched out, and it was simply a whimsical icon added because there was room for it. It didn’t involve Microsoft, either.
But what even is the “Wang word processing character set”, anyway? It took me some digging to find out. Wang Laboratories did produce dedicated word processing systems, as well as word processing software that could run on more general-purpose Wang machines.
Wang Word Processing Characters
In a manual for Wang word processing software I found numerous screen shots showing what Bill Gates was talking about, like this one (from page 5-32 of the manual):
There really are triangles and other interesting characters; these represent carriage returns, tabs, and various control functions. This document (page 69) shows a bit more about how those special characters were utilized by Wang word processors:
The set of special characters was not entirely fixed, although there are clearly many common elements. Here an excerpt from the Wang 2200 Programmer’s Guide to Word Processing (page 1-3):
And this is a look at the character set question from another angle, from a Wang terminal technical reference, a 1981 update of a manual first published in 1979:
Of interest is particularly the 81-8Fh character range (middle column). Those must be the characters Bill Gates was referring to, including left/right triangle, a box, a diamond, double exclamation mark, and several arrows. What is not found in the above character set is any sign of a smiley face character.
But Wait!
For comparison, here’s part of the character set of the IBM PC, from the original 1981 Technical Reference:
Of particular note is the 10-1Fh character range (2nd column of characters). There’s that left and right arrow, a box, a curious double exclamation mark, various arrows, and more. Now, individually those characters aren’t that unique, but I’d argue that as a group they are. Some of the characters, such as the right angle (1Ch) or double exclamation mark (13h) are extremely rarely used.
What are the chances that IBM purely by accident grouped together several unusual characters that just happened to be used by Wang word processors? Certainly not high.
It’s much more likely that there is something to the Bill Gates story, and the IBM PC character set includes those characters precisely because Wang word processors had them. Yet the smiley face wasn’t one of those.
MultiMate
There was a fairly widespread early PC word processor heavily inspired by Wang word processing systems. It was called MultiMate, and although the user interface was clearly meant to look like Wang word processors, MultiMate did not use those Wang-inspired characters, as shown by this table from the MultiMate 3.20 Getting Started brochure:
It’s not at all clear why MultiMate didn’t use the Wang-style characters that were right there in the IBM PC character set.
More Weird Tales
There’s more in the Bill Gates 1995 interview that just does not add up. First Gates says:
There were many issues about this machine. Should it have graphics or not? First design did not. We convinced Eggebrecht that he could do a graphics card. The funny thing was, they wanted to have these good looking characters. So, they did a video card, this character mode card that didn’t support graphics, but they did go ahead and do the color graphics adapter, CGA card.
Then a few paragraphs later, Gates says:
Although we had first just started talking about the BASIC, but then they said they wanted other languages. We knew that we had helped design a machine that was more than a home computer. So, then we were kind of saying, “Well, if you want the other languages, you must know too, this is not just a home computer. Let’s not joke around.” You don’t put COBOL and FORTRAN and all that on this machine. And so that group and Microsoft sort of conspired to make it an all-purpose machine, not just a home machine.
So according to Gates, silly old IBM was designing a home computer, but thanks to Microsoft, it turned into a business machine, too. In other words… we’re asked to believe that IBM was designing a home computer with high-resolution text but no graphics, running FORTRAN and COBOL. That makes no sense whatsoever.
And of course we’re also asked to believe that the IBM team whose most relevant prior experience was the Datamaster, very much a business machine, suddenly decided to throw all that experience out and design a home computer instead. It is worth noting that in 1980, Bill Gates may not have had any idea that the Datamaster even existed, since it wasn’t announced until July 1981 and IBM was notoriously secretive.
What is known is that the PC design drawings dated August 10, 1980 included both a “personal cassette system” with minimal memory and a color TV adapter, and a “business system” with more RAM and a high resolution text-only display. It is highly likely that that was IBM’s plan all along, before Microsoft was involved at all.
There is a possibility that IBM let Bill Gates “convince” them to do what they were already going to do anyway. There’s also a possibility that Gates invented a lot to make himself look good, because he knew IBM wasn’t going to correct him.
There are also statements in that interview that were at best true for only a very brief window of time, and probably not even that, like this claim:
We actually had more people assigned to the project than IBM did. And we were only a company of 30 people.
At the beginning of the PC project, there was already a core group of 12 IBMers assigned to it. We’re asked to believe that half of Microsoft was working on projects related to the IBM PC sometime in summer 1980.
Even if that were true, IBM soon had hundreds of people working on the PC, far more than the total number of Microsoft employees at the time. Once again, the claim is either just plain wrong or intentionally misleading.
What Bill Gates clearly knew was that people who tell a story get to choose how they tell it. IBM said very little, and that gave Gates an opening to put a massive spin on the story of the IBM PC origins, without anyone being able to call him out on it.
Update (Oct 4, 2021): The MS-DOS Encyclopedia claims that “on August 21, 1980, a study group of IBM representatives from Boca Raton, Florida, visited Microsoft. This group, headed by a man named Jack Sams, told Microsoft of IBM’s interest in developing a computer based on a microprocessor. IBM was, however, unsure of microcomputing technology and the microcomputing market.” The purpose of the initial meeting was, according to the Encyclopedia, to inquire whether Microsoft would be able to supply an 8-bit(!) BASIC by April 1981.
The Encyclopedia does not explicitly state but clearly implies that this (August 21, 1980) was the first meeting between IBM and Microsoft. The Encyclopedia also introduces the section by stating that IBM was developing a business computer, not a home computer.
The Encyclopedia very strongly implies that Microsoft essentially talked IBM into building a 16-bit 8086-based machine instead of an 8-bit one. And yet… we now know that IBM already had plans to build an 8088-based PC. It is possible that IBM deliberately played dumb. It is possible Microsoft simply could not fathom that IBM could actually have any experience with building microprocessor-based machines. It is possible that IBM told Microsoft very little, and Microsoft made wrong guesses about what IBM wasn’t telling them.
The Encyclopedia directly contradicts Bill Gates in that according to the Encyclopedia, IBM wasn’t even thinking about a home computer. The Encyclopedia’s account is, all things considered, significantly more credible than the unlikely tale told by Bill Gates.
What makes very little sense is the Encyclopedia’s claim that IBM went to Microsoft to find out whether Microsoft could deliver an 8-bit BASIC within about 6 months. IBM already knew that Microsoft was licensing 8-bit BASIC to numerous OEMs, so unless IBM was thinking of using some exotic CPU, the answer was almost certain to be “yes”.
Now if IBM was looking for 16-bit BASIC, then the relatively long development schedule and high level of uncertainty made much more sense. Given the now-known design plans IBM already had before the August 21, 1980 meeting, there was either some confusion (perhaps 8086 vs. 8088?) or misdirection. The heroic tales of clever little Microsoft telling big dumb IBM how to design the PC are 99% vapor.
The “Lexington” sub-story reminds me of the myth that Winchester drives
were called that way ’cause they were developed in the eponymous town.
Though “Lexington” might indeed have been an informal nickname for the
keyboard in question (among a relatively small count of people).
Thanks for the article and agree with your analysis.
I was employed at IBM for 11 years. When I read the BG quote, I read “Lexington” as a reference to a team at IBM working on the keyboard. We worked together with teams in Lexington, Poughkeepsie etc. and we always referred to the teams/”labs” themselves by those names, since there was usually no ambiguity. By that interpretation, BG is saying that the Lexington lab had made some initial keyboard layout that Microsoft managed to change.
IBM had mentioned a few things about the development of the PC though in rather restrained IBM fashion. The IBM Systems Journal article detailing the development of APL for the PC covers the business case for graphics cards, disagreements with IBM Boca over changing the character set for MDA, standard APL characters replaced with ones from the PC character set, and some other interesting details. Since the project was assigned to IBM Madrid in mid-1981, IBM had a fairly good idea of what the resultant PC should be very early.
That Jan 1985 issue of Systems Journal also has articles on Xenix, PGC, Fortran for the PC, and Virtual Device Interface. It is perhaps the single issue most focused on the IBM PC.
I can see Bill Gates pushing IBM do give a card with both 80 column text and good graphical resolution, basically combining all the display features of all the machines MS had been working on. 80 column mode on CGA was going to happen anyway since IBM needed that.
This is where a more detailed timeline would be great. We know exactly what IBM was planning to build in August 1980. It’s unclear to me when all the sage advice from Bill Gates was supposed to have been dispensed.
I wonder why NEC did not change the switch characters in Japanese DOS. The fact that backslash became an yen would itself be a good reason to do it.
Added more to the article based on the timeline presented in the MS-DOS Encyclopedia.
Interesting. Me wouldn’t be surprised if those initial enquiries were
mere probes in some form.
Smart IBM riding dumb Microsoft?
They were certainly probing how much Microsoft could deliver and under what conditions. BASIC was the obvious starting point since it was built into the ROM (and that was the plan all along).
On the operating system side it’s pretty clear that CP/M-86 was the initial choice, but when DRI couldn’t or wouldn’t deliver, things got interesting. The MS-DOS Encyclopedia’s account is that after Microsoft already committed to supply BASIC, FORTRAN, Pascal, and COBOL, adding a basic OS really wasn’t all that much. Which is actually true, in terms of code size DOS 1.0 was quite small, especially when one subtracts the utilities written by IBM. In fact the core DOS 1.0 was about the same size as BASIC alone.
Jack Sams had done a similar probing for information about a decade earlier to determine how IBM could best support architectural firms which led to IBM buying the design software from Skidmore, Owings, and Merrill*. I am certain that IBM let nothing slip out as to their final plans. I am equally positive that Bill Gates spent his meeting trying to convince IBM to design hardware that would run all of the MS software currently being contemplated including what became Multiplan, Chart, and Word.
Success has a thousand fathers; let’s give Bill Gates his chance to figuratively pass out celebratory cigars.
* See a section of Drawing Futures written by Ann Lui titled “Data Dreams: The Computer Group and Architecture by Spreadsheet, 1967–84.” Interestingly, the piece includes a complaint that “the specifics of the discussion remain unknown.”
Bill Gates also claimed that IBM originally only wanted to support 512K of RAM in the PC and he was the one who personally convinced them to support up to 640K. However, the first two revisions of the PC actually only supported up to 544K of RAM; it wasn’t until the October 1982 BIOS revision and a redesign of the motherboard that it supported up to 640K.
Pretty clearly another self-serving made-up story. The original August ’81 PC Tech Ref says that maximum supported RAM is 256K, which assumes 64K onboard plus three 64K expansion cards. The system memory diagram on page 2-25 clearly shows 256K maximum RAM plus “384K Memory Future Expansion”. 256+384=640, there’s no mention of 512K anywhere.
It’s pretty clear that the architectural limit was always 640K, but in the times of the first PCs there were no memory expansion cards big enough to get anywhere near that. 512K happened to be the maximum supported by many boards of the AT era, but even then it was possible to add a card that filled up the “missing” 128K.
Even with the original PC, the limit came from how the BIOS interpreted the DIP switches indicating the memory size, not any real technical limitation.
Do we all agree on what was considered a “home computer”?
Arguably some of the computers HP made were “home computers” as they had BASIC in ROM even though they were way too expensive for home users and lacked any of the graphics and sound capabilities that a home computer ought to have.
What Microsoft might had convinced IBM to do is to ship the PC with an operating system that loads from disk rather than relying on BASIC in ROM and kind of booting each other language system as if it were it’s own operating system (like UCSD Pascal on for example the Apple II), or for that sake have some kind of operating system in ROM (which would have a file system API and so on but perhaps not a command prompt).
Me’s always perceived designations like “workstation”,
“personal computer”, “home computer”, etc., as being reflective of the
intended role of the machine in question.
A game console is mainly intended for entertainment. A personal computer
is mainly intended for “serious” work. A home computer is a mixture
between the two.
That is, of course, not the whole truth. It’s safe to say that there’s
a spectrum. Some examples–
0. Simon Travaglia wrote the earliest BOFH stories on a Trash-80 (or so
he claimed), at home, only to post them to Usenet at work.
1. The Apple II and C64 straddled the line between home and personal
computers, while the Famicom instead straddled the one between a
game console and a personal computer.
2. The Atari ST filled a similar role, with the TT going more towards
the “personal” end. And then there’s their Transputer affair, the
Abaq — constructed like a home computer, with an ST front-end, but
otherwise an advanced and unusual machine, especially for its time.
3. The Amiga remains an interesting case: it combines workstation and
home computer elements, largely skipping over the “PC” area in
between. (Or perhaps the spectrum wraps around…?)
So, it’s not really that clear-cut. A lot depends on perception.
All kinds of developments since then have made it even more muddy, of
course…
s/game console and a personal computer/game console and a home computer/-1
Also of note is that 6502 was almost never used in pure personal
computers (according to me definition above), while the ST and Amiga had
a distinct home computer aura around them, and were instead based on a
workstation processor (the 68k).
An example meforgot to mention is the Amstrad/Schneider toy line, which
(at least in Europe) was marketed as a serious machine, but which made
a lot of concessions typical of a home computer, with added vendor
lock-in due to the monitor containing the power supply… It basically
had all the disadvantages of a home computer and a personal computer
combined, and then some. (Though others might perceive that
differently.)
BASIC in ROM does not make a home computer. Consider for example the IBM System/23 Datamaster — it had BASIC, but it sure wasn’t a home computer. Or the earlier Wang 2200 — BASIC built in, definitely not a home computer.
[tl;dr version of me 3-post rant]
There’s no real distinguishing feature that makes it “home” or
“personal”, is there…?
Not really. Computers are what users make them.
Then again… an S/360 won’t be a home computer no matter how hard you try.
Heh, as long as there are people with entire railways in their back
yards…
True, but an S/360 would be the equivalent of a full-sized railway in your back yard, not just a model train 🙂 I’m sure there are people who could do that.
Thanks for this article Michal.
I always wondered about the origins of those characters.
You did a great job collecting all the materials and old Wang
screenshots!
There’s nothing like having proven original documents (or often for computer history, either or both of hardware and software/digital evidence) when it comes to being able to at least attempt to sort out what actually happened, and when, in the past! After doing a deep-dive into all the Slack Space I found on what I knew to be an original IBM PC DOS 1.00 distribution diskette (from which I made an image file, and printed MD5 hashes for back in 2005), I was quite happy when Dick Conklin contacted me with an even earlier bit of digital evidence full of original Slack Spaces and earlier versions of what would become the DOS 1.00 code; which we decided to refer to as “DOS 0.90”. Like archaeologists, one can only hope there is still more “evidence” ‘out there’ waiting to be revealed, in order to answer the many questions we still have about the PC history.
RE: Gates… He strikes me as someone who won’t even tell his own kids anything that would ever detract from the picture he’s painted about himself. He did a pretty good job at trying to get geeks to see him as one of them, but the fact is that unlike those who made Apple or DRI (Gary was a sort of military-geek) or many other inventors who became part of the computer industry, Gates’ parents were already Millionaires – when just $1 could buy a lot of stuff back then – and social “influence-rs”! (With some dark history – if you were a baby waiting to be born.) Too few PC geeks ever study the history of the Gates family as a whole; once you do, it does appear to shed some light on what happened between IBM and M$: Although there are no documents I know of, only the memories of certain IBM employees who say they really wondered why they were having meetings with Bill rather than others.