Bill Gates unearths Microsoft's ancient code like a proud nerd dad Microsoft co-founder Bill Gates has shared the 1975 source code for Altair BASIC. The code was the foundation on which Microsoft was built. Before Windows and before Office, there was a carefully crafted BASIC interpreter designed to fit within the limited resources available on the Altair 8800. Why an interpreter? Compiling …
30 years ago as an undergraduate, I remember a Comp Sci professor telling the class that he had a copy of Bill Gates' original BASIC and that it was a really fast, efficient and well-written piece of code. He said Bill is clearly very good at programming.
Then he laughed and said "But I've found a bug in it!"
He didn't elaborate any further. I guess it wasn't a very serious bug.
From Programmers at Work, Microsoft Press, Redmond, WA [1986]:
Interviewer: "Is studying computer science the best way to prepare to be a programmer?"
Gates: "No, the best way to prepare is to write programs, and to study great programs that other people have written. In my case, I went to the garbage cans at the Computer Science Center and I fished out listings of their operating system."
'Nuff said.
If memory serves, he and Allen also used the infrastructure of said Computer Science Centre to write the Altair BASIC code in the first place. Strict interpretation of the rules of use would have made the software the intellectual property of the University, and not Microsoft.
Kids, huh?
Not just kids - grown-ups, too. Tim Paterson of Seattle Computing Products pretty much "created" Q-DOS (Quick and Dirty Operating System) by taking CP/M, filing off the serial numbers and giving it a new paint job, then renaming it to 86-DOS and making it commercially available. Gates bought it, put a Microsoft sticker on the box and sold it as MS-DOS. The similarities between CP/M and MS-DOS 1.0 are painfully obvious.
However, Digital Research (who had created, and owned, CP/M in the first place) never gave permission for their IP to be used in that manner by any third party, which essentially means MS-DOS was released as a stolen product. IBM knew this. They went ahead with it anyway. DR was too small to be considered a threat to Big Blue. When DR threatened with legal action IBM agreed to offer CP/M as an option for the IBM PC next to PC-DOS (the IBM-branded MS-DOS) but ensured that the pricing was such that CP/M was immediately pushed out of the market.
Kids, my rear panel RS-232 port! Gates and Allen stole computer time at Lakeside to develop their own work that they then commercialized, then willingly and deliberately bought and sold stolen property. The rest of Microsoft's history simply continues that trend.
Recent study says 60% of modern Open Source code.copied from other programs... e.h. copy-paste via Stack Overflow.
Worked with chatGPT as a coding assistant?
If the kind of strict interpretation of IP.law were in fashion, the software industry would grind to a sudden halt.
Looks like pretty good code to me. I have familiarity with 8080 code from decades ago, and that helps. Figuring out the meaning of stuff in the assembler, and twisting my brain to octal (been a hex guy for decades) takes some effort. Lots of comments in the code, which is good - way too many "programmers" nowadays either don't comment, or have terrible comments. I haven't spent more than a few minutes, but it looks like the whole thing uses lots of lookups in tables. Uses CPU status flags as character classification - nice!
... is that one way to succeed in business is to lie. "Yes, we've got a BASIC interpreter that'll run on a 4K Altair 8080. No, it's not with me. It's ... in my other coat. Which is at the cleaners."
Mine's the one with the punched paper tape listing of the BASIC language "Star Trek" game in the pocket.
Back in the 1970's outside of the official Dartmouth release and the DEC PDP 11 BASIC compiler if you ran a BASIC program you pretty much always ran on an interpreter. Which is where BASIC got its reputation for being slow. Real slow.
That was the big selling point the PASCAL pushers emphasized - its so much faster than BASIC. Which it was. Up to a point. Then C came along and showed up the weakness of the P-code machine implementation. Most people writing PASCAL "compilers" cheated and pretty much just wrote a P-code interpreter. Which although faster than BASIC was much slower than C. DEC, as usual, did it properly by writing a blindingly fast true PASCAL compiler (with native code generated) with a fantastic optimizer. But C won that war. As you could write a (mostly) K&R C compiler pretty easily that generated good enough native code without too much effort. So by the mid 1980's there were good enough C compilers running on pretty much everything. With PASCAL following BASIC riding into the sunset. Not without some fantastic implementations. Like the original (suppressed by MS) MacBasic. Most of whose features turned up a few years later in Lightspeed C on the Mac.
Borland Turbo Pascal was good too, which evolved into Delphi for writing Windows programs (via an initial release as Object Pascal for Windows, of which the least said the better). They also had sensible pricing for the lower end versions, so even as a student I had a legitimate copy. Then it went to Embarcadaro who turned it into RAD Studio at enterprise-only licence costs. They have a community edition these days, but some surprising features are still enterprise only, such as connecting to a database running on a different machine.
In the 1980s Pascal did quite well, and did it into the 1990s, on PCs. TurboPascal was efficient, fast. had an IDE, and costed far less than a C compiler. It was Windows to move a lot of programmers towards C - since the SDK was for C programmers. Borland Pascal was fully compatible with C ABI, but still header files needed to be translated into Pascal. Then most LoB applications was moved to web ones, and RAD tools like Visual Basic and Delphi became far less useful.
In the 1980s Pascal did quite well, and did it into the 1990s
In the mid-80s I was at Polytechnic (US readers - think community college) doing an HND in IT (1/3 programming, one 1/3 comms, 1/3 analogue electronic [1]). In one of our programming modules, we had to write a simple stock control system. I, being the posessor of a BBC Micro with plenty of (cough) sideways ROM images I could load into sideways RAM, wanted to use the Pascal ROM I had.
"Sure" said the lecturer. "Pascal good."
It wasn't until I started coding it that I realised the BBC Pascal had a number of limitations, the worst of which was that it had no random read/write. So, to update record 120, you had to read the whole previous 119 records, write them to a new (temporary) database, amend the 120th record, append it and the rest of the records to the temporary file then replace the live file with the new one.
It (kind of) worked but, being entirely floppy-drive based, was extremely slow. Slower than treacle at midwinter.
So I junked it and just wrote a convincing demo instead. And got a pass grade because I'd shown "inititive and the ability to code, even if I didn't actually deliver the requirement".
Which taught be a lot about the IT industry.
[1] I *hated* analogue electronics and it's why I, eventually, dropped out of the course, despite getting really high marks in the other two sections.
You're being a bit hard on Pascal implementations there. It is is fair point that a lot of microcomputer compiled languages used some sort of p-code (UCSD Pascal most obviously), and I am sure there were others, if only because generating good re-entrant stack based code is a challenge on a Z80 and a nightmare on the 6502. The 6809 was probably the least awful 8-bit platform for that sort of native code. But true compilers were the norm in the minicomputer world. I remember working as a Southampton Uni undergarduate with OMSI Pascal in 1981 on a heavily overloaded PDP-11/45 running RSTS/E. A year or so before. Boffins in comp. sci. had ported Pascal on some ICL 1900 thing and a Honeywell doo-daa that were the mainstays of campus timeshared computing at that time. I think at the end of the day, too many people were seduced by C's ability to get down to the metal and relative flexibility on type conversion, in an age where programmers still had to consider hardware limits on 16 bit microprocessors. But that's a religious war for another time, I suspect.
In Pascal 6000, at least, you could use non-discriminated type unions to place whatever interpretation you liked upon a variable or structure. They were little-talked about or documented, and considered as like to necromancy, because if you erred, Bad Things would happen within your program. Further, they were not portable. But they worked great for me the one time I had to use them, to debug a recursive program with pointer problems.
I have a friend who's a bit of a novice and wondered why his shiny development M1 laptop was thrashing hard. On closer inspection it turned out this fairly recent machine had 8gb ram in it. Explaining that I had double that on my GPU alone, on a rig that cost half the price, a lesson was learned about the cost of poor hardware choices.
Well, he sold the source code of a game called Bastar in BASIC for the SpectraVideo SV-318 micro computer to a US computer magazine for publication. Beyond that, probably not much from him around. It is even questionable if he even did any coding working himself at his first company, Zip2, which he founded with his brother Kimbal and a guy call Greg Kouri, who put up the first money and probably did most, if not all of the programming for their "online city guide"...
That paper tape / printout that cause quite a kerfuffle back in 1975. That's when we first knew that Bill G was a complete and total conniving bastard.
The story was that like so many of Bill G's future products it was mostly a rehash of other peoples work. There were a bunch of other Tiny BASICS doing the rounds back then and the story was that quite a bit of BillG et al's 8080A code looked remarkably like some of the other interpreters out there at the time. Like line for line translations of blocks of code. But hey, Microsoft would never ever do anything like that. Would they.
https://en.wikipedia.org/wiki/An_Open_Letter_to_Hobbyists
In 1975 I was chugging along on DEC's Tiny BASIC for the PDP-8E. Which ran useable BASIC programs in 4K . Of 12 bit core. But ran better on the teletype which had 12K allocated to it Heady days.
"Like line for line translations of blocks of code. But hey, Microsoft would never ever do anything like that. Would they."
I'm somewhat curious why a purportedly 70-ish-year-old geezer still decides to post anonymously instead of his forum handle.
Can you substantiate your claims now that the source code is published?
You obviously were not around at the time.
If you were around in 1975 to 1978 in the very early days of the S-100 based and all the other early micro-computers you would have been very familiar with all the 8080A (then 6502 / 1802 etc ) software listings that were kicking around. In fact early issues of BYTE / Creative Computing etc had ads for books of these listings. Most were for BASIC but a whole bunch were for assembly language software. The asm listings were mostly for monitors, hardware interface and such like but there were at least two were for Tiny BASIC interpreter listings from what I remember. Even flicked through one of them at the time. Theses books were just 100 / 200 pages of print-outs. On paper. Which had to be hand inputted. Typed in. Or even keyed on on the smaller systems. Some of these books even had proper covers. And were not stapled together. Or ring bound.
Then there were all the big stapled together xeror pages (or just fan-fold) listings in circulation. There were plenty of those kicking around. In fact I threw out the last ones I had about a decades ago. For the 1802. All the thick fan-fold listings had been throw out decades ago.
Almost all the early microcomputers had no floppy drives (let alone harddrives ). You might have cassette tape as permanent storage. Some might have repurposed mini-computer paper-tape readers. Not many. Most micros which could read and write paper tape used surplus teletypes. Look at the early issues of BYTE and Creative Computing. Then Kilobaud a few years later. From 1975 to 1977. Look at the ads. You wanted storage better than a cassette it as going to cost you $K's (of 1977 dollars). Although people like Jameco were selling surplus mini computer floppy drives which might work with a S-100.
Then the Apple II started shipping in mid 1977 and the scene quickly changed. Although it was not until the first Apple II floppy drive shipped mid 1978 that relatively affordable microcomputers started having the same "storage" capabilities of the ten year old plus PDP-8E I had been using up to then. A big rack. A few ASR 33's. And a PDP-11 high speed punch/reader. In fact with the first Apple II's I played with at the end of 1977 came the listing for "SWEET 16" which was the key bit of the Integer BASIC interpreter that Woz wrote for the Apple II. Which actually worked, was clean and fast. Unlike the other BASIC interpreter shipped with the early APPLE II's. From MS. On a floppy. Which was a "complete" BASIC, very slow, used up almost all memory and had a tendency to lock up.
Its all in the Internet Archive. All the early issues of the US microcomputer magazines. From 1975. By the time the first UK microcomputer magazine came along in early 1978 (PCW) this era was pretty much over. But early issues were full of ASM listings Including a monitor written in Z-80.
So yeah. I know what I am talking about. Been in from the very beginning. And worked in the business in the US for so long that I remember when the (now mostly gone) buildings on Mariani Ave was the height of Apple affluence and the area around Bel / Red, 148'th to 156'th and over to 520 (and now beyond) on the Eastside was still mostly undeveloped. Not Borgville
As for any animus towards BillG. It's just professional. He was a crook who should have went to jail long ago. But when he was out and about in Madison Park always did the Seattle Nice thing and just nodded, smiled and said hi.
Some of us actually work in the business. And still do. And know were all the bodies are buried. And who put them there.
That's a big wall of text not related to my question at all.
Since the Altair BASIC code is now out: can you or can you not point out the "line for line translations of blocks of code"?
"He was a crook who should have went to jail long ago."
For the alleged copying of BASIC?
"Some of us actually work in the business."
Sad attempt at belittling me. I'd say most of the active Reg forum dwellers are in the business, moi included.
@Sandtitz
That "wall of text" is a detailed description of what was going on in the mid to late 1970's in the micro-computer / early personal computer scene. From someone who was there. From the very beginning. Is it beyond your reading comprehension level perhaps?
So I take it you never worked in the commercial personal computer software dev business. Shrink-wrap software. Have zero experience writing / shipping microcomputer / personal computer software anytime over the decades. And based on the fact that some very deliberately business insider references I made in my explanation seemed to have gone completely above your head you are just some civilian who has absolutely no direct experience of the personal computer software business, as it was, and how it developed, in the US. In the Bay Area and in Seattle. If you had actually worked in the business, the one that BillG et all were in, you would have got the insider references. All of them.
"Prove it"? FFS. Are you some a*hole troll on The Well circa 1990? Thats the stupid asshat stuff tried back then. You sound like some enterprise IT type or script kiddie who thinks that because he writes some shell scripts etc and goes to trade shows he "works in the business". Sorry to disabuse you. Your professional background (based on how little you could comprehend my detailed technical explanation of what happened back then) bears as much relationship to my end of the professional business as a guy who makes wedding videos in somewhere like Wigan does to a first unit director in Hollywood for the movie business.
Why do I suspect you probably dont know enough about either business to get that reference either.
As for BillG a criminal. Well thats what the anti-trust case was all about. Which they lost. And why BillG was eased out. But that another story. One I am certain you would never understand if explained to you. Never having worked in the business. How about the defrauding the state of Washington of billion of sales tax. Another of BillG's finest moments. One of many. So so many.
"Is it beyond your reading comprehension level perhaps?"
Not at all. Maybe you were there or not. To me you're just another AC here. Maybe I am the Postmaster General.
I asked for references to "line for line translations of blocks of code" as you put it.
I can now see that you are not going to address that part and you're just bullshitting your way out. HAND.
> In fact with the first Apple II's I played with at the end of 1977 came the listing for "SWEET 16" which was the key bit of the Integer BASIC interpreter that Woz wrote for the Apple II.
SWEET16, although present in the ROM, isn't used by Wozniak's BASIC.
> Which actually worked, was clean and fast. Unlike the other BASIC interpreter shipped with the early APPLE II's. From MS. On a floppy. Which was a "complete" BASIC, very slow, used up almost all memory and had a tendency to lock up.
Appelsoft, as Microsoft's BASIC for the Apple II was branded, wasn't known to be slow or to lock up at all. That's probably why it replaced Wozniak's BASIC as the ROM-resident one as of the Apple II+, i.e. starting in 1978. Being an instance of Microsoft's 6502 BASIC it's very similar in speed and never-locking-up to BASIC on a Commodore, Oric, etc.
@ThomH
As yes, you. If I remember correctly someone who thinks they know all about the Mac 64K ROM because they did some minor hacks to get ROM images up on running on hardware emulators. Just like people were doing back in the late 1980's. On actual hardware. On Amigas. Whereas some people actual disassembled and documented the (non QD low level) 64KROM back in 1984...
I have (had?) Woz's IntegerBASIC listings somewhere. Its good 40+ years since I last looked or cared what exactly did what. But the SWEET16 listings were in the docs we got late 1977. Had a quick look and had better things to do. Once heard it was used in the IntegerBASIC. Which kinda made sense for an interpreter. As SWEET16 was an assembler. I still have a stack of the original Apple II (not II Plus/IIe) manuals. Somewhere. But honestly cannot be arsed to dig them out.
So I take you were programming 16K Apple II's in 1978 trying to get stuff working in AppleSoft BASIC because the docs said it had all these advanced feature. Then gave up because it was so damn slow and a lot of the commands did not work. Or just locked the machine. Which mean reloading the whole damn thing from the cassette. Which took a long time. So used Woz's IntegerBASIC to get work finished. And did for the next few years. Until MS finally got it stable around 1982/83. Story sound familiar?
What was shipped with the Apple IIe by 1983/1984 (5 years later) was a totally different product. All the important BASIC commands worked (good enough) and did what you expected but it was still slow. Which is why there was a healthy market for non MS BASICS on the Apple II. Look at the ads in any issue of BYTE for 1982/1983.
Ever ship commercial software for the Apple II? Which people paid for. I did. In 1983. For the the IIe. It also ran on MS-DOS. In a cut down BASIC. Which ran on the bundled MS and other third parry BASIC's. My first and last BASIC product. After that it was C or ASM all the way.
So yeah, none of this is purely theoretical or second hand knowledge / opinions. Which I'm afraid your postings mostly seem to be.
You are making assertions based on third party opinions of people who were either not there at the time, did not use the products when they were first released, or never tried to use them to write non-trivial commercial programs. During the commercial lifetime of the hardware or software products.
Proof if any were needed that age does not imply maturity.
> You are making assertions based on third party opinions of people who were either not there at the time, did not use the products when they were first released
It is a stone cold fact that SWEET16 is not used by Woz's BASIC. This is not an opinion. It is an objective reality. It speaks to the frailty of your intellect that you don't understand that facts are not opinions, though it does somewhat explain your negligible grasp of facts.
The original Apple ROMs are a collection of things Woz thought would be useful. The original integer BASIC is one. SWEET16 is another. Floating point routines are a third. The monitor and mini-assembler are others. The use case was just that anybody with an Apple II might want to use any one of these things.
The norm of 8-bit computers booting into BASIC (in many cases being pure raw-metal implementations of BASIC rather than anything more layered) is a later thing. You shouldn't view what's in the Apple II ROMs are being coupled to what is needed or used by its original BASIC.
There were plenty of listings kicking around in the mid 1970's. Including for BASIC interpreters. In fact that's how we learned this stuff. By reading the listings.
You wanted to learn how to write an interpreter or compiler. Find some listings. And read very carefully. The first useable book published on the subject (for non-functional languages) did not come out till the mid 1980's. All previous books were purely academic courseware unusable for writing real world compilers or interpreters. The first useable book for writing functional language compilers did not come out for another decade.
So yes, there were plenty of listings available. And that's how this stuff was written back then. Read other peoples work and reimplement. Including BillG's minor opus. Read the DEC10/20, PDP-11, IBM 370 etc assembly listings. Translate to 8080A. And then try to get all that to fit in a few K. Now that bit did take some skill. But as Woz proved a two years later with the Apple II Integer BASIC BillG and Paul Allen were just not in the same processional software engineering league as people like Woz. Not even close.
Now Woz, he did some truly awe-inspiring work. Even now looking over the original Apple II motherboard schematic, compared to all the other 6502/8080A boards at the time, is just a work on inspired genius. It's in the back of one of the manuals that shipped with the Apple II. The schematic for Woz's floppy disk controller was in the back of the Apple II floppy disk manual.
A very different world.
The established timeline is that the Tiny BASICs were a response to Altair BASIC and the open letter, not that Altair BASIC took from the Tiny BASICs.
So you'd have to allege that Altair BASIC took from previous BASICs, none of which were written for the 8080, and billg stitched Altair BASIC together from sections of code written for other CPUs. Is that what you're alleging?
If so, which BASICs and where is the source code to those BASICs?
Wah wah wah, I'm old and I'm old and BillG keeps hanging around my bins and the Apple II had HDMI and if you weren't there then you don't even know that Micro$ucks and the only reason you're even asking that is the 5G in the COVID.
It's just sad the way that Windows 3.1 was such a rip-off of OS X 10.3, which has been proven by all the best guys and is a real fact. These guys come up to me crying and they say, sir, OS X 10.3, from 2003, was so unfairly treated when the 80286 came out. Big guys. Like Larry Grayson, who runs Oracle.
AC:
There were plenty of listings kicking around in the mid 1970's. Including for BASIC interpreters. In fact that's how we learned this stuff. By reading the listings.
Cow:
Well the first author of the first basic obbviously had to invent everything themselves, so that isnt true, some people do figure things out themselves.
AC: Read the DEC10/20, PDP-11, IBM 370 etc assembly listings. Translate to 8080A
Cow:
Those CPUS work very differentl to the 8080A, completely different register sets, with completely differently addressing modes and op codes.
Given the very limited ROM space, it would be terribly inefficient to try and emulate a basic from those other cpus, simply because one opcode from there would become too many in 8080A, and there would be a lot of memory loads and stores which would make it incredibly slow. You can tell by reading the 8080A basic its been hand crafted to use fit the limitations of how a 8080A worked.
Say what you like about BG and Microsoft - but that's an incredible ballsy move - not only to promise you have a working application when you don't, but to sell someone an app for which you've not yet written on a platform you don't own and can't afford to buy. Solution? Step 1 - write an emulator for the platform.. Step 2 - write the app on the emulator.. Step 3 - delivery product in under 8 weeks. Talk about agile - some times it takes my work 8 weeks to do the paperwork needed to approve the paperwork needed to start work on any code.
I definitely remember attending interviews (1980?) for Comp Sci course at Queen Mary College, London. Reading the interviewer's form upside down on his desk, he had a checkbox marked "BASIC Damage" such was the bad reputation that BASIC had in academia. Seemed a bit harsh at the time, though I understand the point.
The PDP-10 (&11) were usually termed mini-computers not mainframes (limited address space). It still looked like a blue commercial freezer. I show my age by having actually programmed one. I know someone who hollowed out the gubbins of one and used it as a clothes cupboard (nerd award).
Its a long time ago (1972'ish?) but IIRC, the Intel 4004 was originally advertised as 'a microcomputer'.
(The image used in an Intel advert of the time was that of a few 16pin DIL packages (4004, along with 1, 2 & 3) standing up on end in a sermi-circle and intended to look like a 'normal' minicomputer's bays).
If thats all bollocks, just put it down to an old bloke's memory rot.
This post has been deleted by its author
Pre dot-matrix lineprinter fan-fold paper does bring back the memories. Modern development tools obfuscate the liniear nature of a programme and the (correct) order in which things need to be declared, although a screen full of code is probably as good as a page full of code in encouraging good segmentation of logic.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
