Happy 30th Birthday, PC
If your motherboard has an old-school BIOS, rather than a swanky new EFI system, then try counting the number of character columns across the screen next time you’re fiddling in it. Alternatively, if you have better things to do than counting letters, and we assume that you do, then take our word for it that there are 80 columns. The reason for this is that there were also 80 columns on a standard punch card – the archaic paper-based storage media of early computers that was invented by Herman Hollerith in the 1880s.
Along with many other parts of the PC architecture, this is a vestigial remnant of the original IBM PC 5150 from 1981; the machine that bridged a gap between the computing relics of yesterday with the PC as we know it today. ‘All green-screen terminals had 80 columns,’ explains Terry Muldoon, former IBM systems engineer (and now the museum curator at Hursley, following his retirement), as he shows us round the archive of relics at IBM’s UK headquarters in Hursley ‘because they were basically emulating a punch card.’
The IBM PC 5150, the world’s first PC, kickstarted the computing revolution 30 years ago today
Computing was radically different before the advent of the personal computer. Large businesses with a computer system would generally have a mainframe computer, which would then connect to operator consoles, or terminals; the idea of having a whole computing system in front of you on your desk was an alien concept.
The Xerox Alto had a GUI-based interface,
despite being built in 1973
This isn’t to say that there weren’t experiments with personal computers before the IBM PC 5150. The future-looking technology gurus at Xerox PARC, for example, developed the Alto personal computer as far back as 1973, complete with a GUI-driven operating system.
However, only a few thousand were produced, as it wasn’t practical to mass-produce in a cost-effective manner. Likewise, IBM had dabbled with the idea of computers that were independent from a mainframe with machines such as the 5100 and 5110, but again the high cost made them unattractive to the mass market.
Meanwhile, Commodore had similarly experimented with the personal computer concept when it released the PET in 1977, as had Apple with the first kit-based Apple I, and later the Apple II. However, the market for the personal computer took a major turn into the mainstream world in 1981 when IBM launched the PC 5150.
‘This is personal opinion,’ says Muldoon, ‘but I think someone in IBM saw the rise of the personal computer when they saw machines such as the Apple IIC running VisiCalc. We had products such as ADRS, which was basically a mainframe spreadsheet, and I think they saw that this was often more capably done on a personal computer. We didn’t think we were going to sell many though. I remember the first year we sold 50,000 PCs in the UK, and we all said “Yes!” – we couldn’t believe it.’
IBM PC 5150 Up Close and Personal
The PC 5150 is remarkably primitive compared with modern standards. The machine had just 64KB of RAM as standard, which could be expanded to 256KB via 64KB expansion cards. There was also no hard disk, with storage duties handled by a 5.25in 360KB floppy drive. Meanwhile, depending on your monitor and graphics hardware, you had a choice of a monochrome text-based character display or CGA graphics, the latter of which could display only four colours at a resolution of 320 x 200.
Also of note is the processor – a 4.77MHz Intel 8088, along with a socket in which you could install an optional 8087 maths co-processor for hardware floating-point mathematics. The processor was an interesting choice though. The PC architecture has rigidly stuck by Intel’s x86 instruction set ever since, but this could have been very different if IBM had chosen to use another type of processor.
In particular, IBM spent a lot of time developing its own RISC-based processor called the 801. ‘If I remember rightly, the 801 was actually looked at for the 5150, or the follow-on from the 801,’ says Muldoon, ‘but I’m not sure there was a single-chip implementation of the 801 at the time.’
Click to enlarge
The single-chip design of Intel’s CPUs wasn’t the only feature in their favour though. Remember that IBM had already had experience with the 8085 in the System/23 Datamaster, and the 8086 in the DisplayWriter – a standalone word processor that was 8086-based, with a screen that’s almost identical to a monochrome PC screen, as well as floppy drives, a printer and so on.
Experience with the Intel architecture was a key factor. ‘The original PC is absolutely keyed to the Intel bus and Intel interface,’ explains Muldoon. ‘The address bus, the data bus and the control bus are just extensions of the processor, and IBM already had experience doing that with Intel’s buses with those other two products, so I think that’s why IBM went with Intel for the 5150: experience.’
Remember DOS?As well as the hardware, the PC had to have a self-contained operating system of its own, and this presented a problem. ‘IBM didn’t have time to write one,’ explains Muldoon, ‘so it went out to various people, including Digital Research for CP/M. Digital Research didn’t return the call, but Bill Gates did.
‘Gates didn’t have an operating system, so he bought QDOS and then went up to IBM and said, here you are – at least, that’s the story.’
‘The original DOS was a tarted-up QDOS, supplied to IBM, so there was IBM Personal Computer DOS.
‘Gates was allowed to sell Microsoft DOS (MS-DOS), and it carried on for many years with exactly the same numbers. So [version] 1.1 was DOS 1 but with support for us foreigners, then we went to DOS 2 with support for hard disks, DOS 2.1 for the Junior, and later the Portable PC, as they had different disk drives and so on.
Muldoon’s personal opinion is that allowing Gates to retain the intellectual property for the operating system was a mistake. ‘So we now have an Intel processor, which is tied to Intel, and we have another guy who owns the operating system,’ he points out, ‘so you’ve already lost control of your whole machine around 1981 – the rest is history.’
IBM’s senior sales consultant Tikiri Wanduragala, who started at IBM around the time that it introduced the PC XT, sheds some more light on this situation, however. ‘At this same time, there were anti-trust cases going on,’ he explains.
‘You can’t make decisions where you appear to be doing certain things. So it’s easy to say that was a mistake, and this was a mistake,’ he continues, ‘but trying to own everything on a product wasn’t the smartest thing to do at that point. Business decisions went wider than just the product – at that time we had a very large number of shares in Intel, for example.’
What Happened to the 186?
There’s a glaring omission in the PC’s processor history during the 1980s – the 80186 processor, or 186 for short. We move from the 8088 and 8086 to the 286 – why wasn’t there a 186? In fact, Intel did create a 186 processor, which introduced more instructions to the architecture. However, it wasn’t compatible with the IBM PC architecture, so it was only used in a few embedded devices.
Interestingly, ex-IBM veteran Terry Muldoon says he heard an amusing story, which may or may not be entirely true, about why the 186 wasn’t compatible with the PC. ‘When it was designing the PC, IBM used some of the reserved Intel interrupts,’ he explains.
‘However, when Intel was developing the 186, the IBM rep wasn’t there on the day that they discussed the interrupts, and they then used up some of their reserve interrupts that were also used by the IBM PC. So although it became a good embedded processor, you couldn’t use it for the IBM PC, since there would have been an interrupt clash.’
There was one crucial aspect of the PC that IBM did own, however, and that was the BIOS. As well as being responsible for controlling the hardware, the BIOS also had the ability to detect new hardware. ‘At POST time, the BIOS goes through and looks for addresses,’ explains Muldoon, ‘and if it finds executable code at those addresses, it executes it and makes it part of itself. So when IBM brought out the XT, the BIOS code for the disk drive was actually on the adaptor card, so when it booted up, it saw the address of that adaptor card, executed it, made it a part of itself and then supported a hard drive.’
The BIOS was an essential part of the PC, which meant that you also needed a BIOS of your own if you wanted to build a fully IBM PC-compatible clone machine. Muldoon gives the example of the software interrupts in the BIOS, including a timer tick, which was often required by IBM PC software. You couldn’t create a clone PC without a BIOS, and this presented a problem for third-party manufacturers.
It’s only recently that the 30 year-old BIOS has been replaced.
These companies had two options at their disposal. They could copy IBM’s BIOS and hope that IBM didn’t notice, or they could create a ‘clean room’ BIOS from scratch that circumvented IBM’s patent.
The latter was a popular option that was employed by Compaq when it created its first Portable Computer, as well as Phoenix when it made the first mass-marketable third-party PC BIOS. ‘One group of people explains what it does,’ says Muldoon, ‘and another group of people takes that specification, don’t talk to them, and write some code to make it do that. That’s a clean room – in other words, nobody has copied IBM code, and there’s a Chinese wall between these two people.’
Muldoon notes that Phoenix did indeed create a genuine clean room BIOS from scratch, but that other clone manufacturers weren’t so diligent in their efforts. ‘Because we published the BIOS, they just copied it,’ he says. ‘Now the BIOS had bugs in it, and we knew they’d copied our BIOS because they’d copied the bugs as well.’
As IBM owned the patent on the BIOS, and it didn’t provide licences to copy it, Muldoon says the theory was that clone manufacturers with a copied BIOS owed the company money. ‘There was a whole group in the UK who went after people,’ he says, ‘and they’d knock on their door, quite politely, and say “I’m from IBM, would you tell me how many computers you’ve made, because you owe us money.”’
Surprisingly, you can’t just plug in a PC 5150 keyboard and expect it to work on a new PC, although it doesn’t seem as though it would be an issue. After all, the 5-pin DIN plug on the end looks identical to the connectors for AT keyboards, which can easily be plugged into a standard PS/2 port with an adaptor. However, the original IBM PC 5150 (and the later XT 5160) used different keyboard protocols from later keyboards, so they won’t work with an AT-to-PS/2 adaptor either.
Click to enlarge
This compatibility issue was widely known throughout the 1980s, and many companies made switching keyboards that could be used in either XT or AT mode. However, finding one of these so that we could fire up our keyboard-less 5150 proved to be hard work, until we called keyboard manufacturer Cherry.
After all this time, Cherry still makes an XT/AT switching keyboard (called the G81-1860), so you can theoretically use it on both an IBM PC 5150 or a new PC via an adaptor. The company also very kindly loaned us a G80-1000 keyboard from its museum in Germany, so that we could get our machine working. It’s testament to the huge popularity of the original IBM PC 5150 and 5160 with businesses that, after 30 years, you can still order a keyboard for them.
It’s clear that the PC has had a major impact on computing at businesses, but there’s no doubt that it’s also transformed the concept of a home computer. Family computers are no longer underpowered 8-bit boxes, but full-on PCs. We point out to Muldoon that the PC is now considered the cutting-edge gaming machine, and PC gaming is a huge industry. Would he or anyone at IBM ever have envisaged this happening?
‘No, not in a million years,’ he says frankly. ‘The games were crude and rubbish. We had 101 monochrome, which actually didn’t have graphics, it was all character-based. If you look at the character set, there are lots of graphics, so you can do boxes and all that
kind of stuff.’
Early PC games, such as Batalia, used ASCII to render graphics
‘Then we had Flight Simulator, which was appallingly bad. It used to run at something like 320 x 200 with four colours, and it was slow. Then we had adventure games, and because of the technology we were using at the time, a colour graphics adaptor (CGA) usually had 16KB of memory on it – you can’t get much into 16KB, no matter how clever you are with your graphics.’
You could get games for the PC, but the limits of CGA meant that they looked primitive – even the ZX Spectrum could display more colours in its graphics mode. After CGA came EGA in 1984, offering 16 colours at a resolution of 640 x 350. However, the cost of an EGA card, as well as a brand-new monitor, ran into thousands of pounds.
‘It was for genuine enthusiasts,’ recalls Muldoon, ‘so it was too expensive to use for gaming. PCs were always just a little too expensive for gaming until VGA.’ VGA stands for video graphic array, and it enabled a resolution of 640 x 480, as well as up to 256 colours, depending on the spec of the card. However, many 8-bit VGA cards enabled the extra resolution but could only display 16 colours.
PC gaming has come a long way in 30 years – The Witcher 2
It was only when VGA became a standard, inexpensive inclusion on every PC that it could finally match home computers such as the Commodore Amiga and Atari ST in terms of gaming capabilities. This isn’t to say that there weren’t gamers using PCs in their own homes though.
By 1987, machines such as Amstrad’s £399 PC1512 started to become popular as family computers that could also be used for work. These machines were primitive – often featuring an 8MHz 8086, 512KB of RAM and CGA graphics. However, they could also run adventure games, as well as a number of arcade titles – you just had to be prepared to tolerate a terrible colour palette.
IBM Makes Its Exit
Not only did graphics cards become affordable, but the popularity of cheap clone machines, as well as improvements in silicon fabrication processes and manufacturing, also dramatically reduced the cost of the PC.
The end result was a computer standard that was surprisingly successful. ‘Who’d have thought that everybody could have a word processor?’ asks Muldoon. ‘Before, these were very expensive machines – the ones with the inkjet printers were around £20,000.’ The declining cost of PCs has also amazed Wanduragala. ‘What amazes me is how the price has collapsed – we’d never have guessed that PCs would be available in Tesco.’
IBM continued building PCs until 2004, when it eventually sold off its PC and laptop divisions to Lenovo in China. Why did IBM abandon the hugely successful product that it was responsible for creating? ‘I was really saying we should have got out of PCs ten years earlier,’ admits Muldoon, ‘it patently wasn’t making money. It’s just difficult to make money in a commoditised market. It requires a degree of skill, which is different to the skill sets in IBM.’
Put simply, the skill sets in IBM are much more geared towards intellectual property and innovation than physical products. The PC isn’t the only product that it’s later abandoned – the same thing happened with punch card machines, typewriters and delicatessen scales.
Back in the early days, AMD simply made licensed copies of Intel CPUs, such as the 8088-clone
‘In the 1960s, Tom Watson Junior said that he’d been chief executive for ten years and IBM wasn’t selling one product that my father sold,’ says Muldoon. ‘So, within ten years, he’d turned the company from a punch card company into a computer company, and now that continues.
‘IBM doesn’t make products – it’s more of a machine that wanders through history, picking up technology, developing technology and then using it to do exactly what Tom’s father was trying to do when he joined in 1914, which is to apply the technology to the customer. The key differences between the company now and then,’ he continues, ‘is that in 1914 it would build a machine and a salesman would go around and say, do you want one of these? Now they go to the customer and say, what are you trying to do?’
Will IBM ever build another personal computer? ‘IBM will make something,’ says Muldoon, ‘but I bet it won’t be a personal computer. It will be something of a groundbreaking nature. You know, IBM invented the mainframe computer, bar codes and so on. It won’t be a PC, and it won’t be a mobile phone because, frankly, LG, HTC and Samsung do it better – and there’s no innovation in it. What IBM will do is stitch these together – that’s where the innovation is.’
Going full circle
The irony of the situation is that after revolutionising the PC, computing is now increasingly turning towards cloud computing, a system that has more in common with mainframes and terminals than with PCs.
‘It’s very cyclical,’ says Muldoon. ’I’ve seen it go from terminal, to standalone PC, to PC connected to a mainframe, to PC connected to a server and then coming back to PCs. The cloud is just that cycle coming around again. It’s all about connectivity, in that there’s the cycle of the personal computer standing alone, and then you get more power from somewhere else, connect it and so on.’
Click to enlarge
There is, however, one crucial difference between mainframes and the cloud, other than the fact that one is online and one isn’t.
‘What’s different from the old terminals is data creation,’ points out Wanduragala. ‘Data is now created on the edge in huge quantities, which didn’t happen with the terminal devices. The processing is in the centre, but the data creation – people taking pictures, audio files and so on – is happening on the edge.’
Bit-tech would like to thank the following people for their help with this article: Tim Beattie for the loan of the IBM PC 5150; Terry Muldoon, Tikiri Wanduragala and Anne Kristine Janson at IBM for showing us the archives and answering our questions; Robin Bithrey at Cherry for quickly helping out with the loan of an XT keyboard.
IBM shows us some archaic artefacts from the early days of personal computing.
Don’t be fooled by the comparatively small CRT monitor; the actual computer behind this would have been colossal.
This is the face of personal computing in the 1970s, when your computer was a terminal that connected to a mainframe computer.
Former IBM veteran Terry Muldoon recalls a client that sang the praises of this terminal. ‘The guy just pointed at it and said “this is so good!”‘ says Muldoon.
‘Apparently there was a half-day course to learn how to use this, and a graphic utility, so he could now do business graphics in front of him and print it out. The trouble was you needed a mainframe at the back of it.’
The cost of the mainframe? Several million pounds.
This is more like it, you might think. There’s now a screen built into a computer, as well as disk drives and a printer. Meet the IBM 5110; the beginning of what you could loosely term ‘portable computing.’ Announced in 1978, the 5110 was the successor to the earlier 5100, adding support for floppy drives. Two 8in floppy drives can be seen in the huge tower box, as well as an 8in disk itself, with a box of Ibuprofen for scale. Despite their huge size, 8in floppy disks from this time were only single-sided, single-density, and could only hold 160KB of data.
Click to enlarge
Even at this point, however, the screen is really only a small part of the equation. ‘People ask me why there’s a printer there,’ says Muldoon, ‘and it’s because this is really an operator’s console. You set the program running, and the program prints out – you’re using this screen to control the machine, but the program is on the printer.’
Click to enlarge
IBM System/23 Datamaster
Moving one step closer to the PC as we know it is the System/23 Datamaster from early 1981. Based on an Intel 8085 processor, the machine can address up to 256KB of memory and features a pair of 8in floppy drives.
However, the machine still cost $10,000, which meant it was quickly usurped by the significantly cheaper (and generally more powerful) PC 5150 when it was released in the US later that year.
After the PC 5150 came the IBM PC XT 5160 in 1983 – the same year that the PC 5150 was finally released in the UK. The PC XT included a hard drive as standard, as well as MS-DOS 2 to support the hard drive and directory systems for file management.
The PC XT could also address up to 640KB of memory, rather than the 256KB limit on the 5150.
IBM PC Jr 4860
In order to make the IBM PC more appealing to the home market, IBM released the PC Jr in the US in 1984. The standard spec came with the same 4.77MHz 8088 CPU as the PC 5150, along with 64KB of memory. However, the machine also came with a pair of joystick ports as standard, as well as ROM cartridge slots in the front, which could hold games. The machine was a major flop in the US, and as a result was never released in Europe.
Portable computing took another step forward in 1986 when IBM released the Convertible, which was also its first PC to feature 3.5in 720KB disk drives, rather than 5.25in 360KB drives.
The machine is called a Convertible because you can detach the screen and use the base unit as a standard desktop machine under a monitor, or you can fold it up into a portable unit with handle that weighs 5.8kg.
My First PC
The bit-tech team reveals their first experience of one of the most important inventions of the last 30 years. Be sure to let us know about your first PC in the comment thread.
James Gorbold, Editor
First PC: IBM PC 5150
I was lucky growing up, as my father worked at the cutting edge of the electronics industry during the 1970s and 1980s, designing test and measurement devices. This meant that I was exposed to some of the first PCs, namely the Commodore PET and the now more famous IBM 5150, almost a decade before most other people in the UK.
However, while my father used these PCs for his job, I was more interested in playing games that were rendered in ASCII such as Batalia and Ikebukuro; I later learned to write my own text adventures in BASIC. Seeing the working IBM 5150 that we photographed for this feature was an emotional trip down memory lane. If I hadn’t had access to these PCs all those years ago, I may not have become an IT journalist and later the editor of bit-tech.
Clive Webster, Deputy Editor
First PC: 400MHz AMD K6-2
Towards the end of the last millennium, my first PC was, on paper at least, awesome. It had a 400MHz AMD K6-II processor, a 10GB hard disk and an ATI Rage Maxx GPU. However, as the PC ran Windows 95, it was constantly breaking down and the more I researched fixes to the various problems, the more I discovered that headline specs don’t tell the whole story. For example, the awesome-sounding graphics chip was integrated onto the motherboard and was the first component to be retired (in favour of a £300 GeForce 3 card). Despite the research I’d carried out before advising my parents about which PC to buy, we ended up with a PC that didn’t last long. All my tinkering and fixing gave me enough knowledge to realise that if I wanted a fast PC that would last, I’d have to build it myself. Eleven years on, I’m still doing exactly that.
Ben Hardwidge, Features Editor
First PC: Sinclair PC200
The first PC I used was an Apricot F1 clone machine that my dad kept for work in 1984, but the first PC we had at home was a Sinclair PC200 in 1989. It was awful by modern standards, with an 8MHz 8086 CPU, 512KB of RAM, no hard disk and a terrible CGA graphics card, which ran all my graphical games at 320 x 200 with four colours. Magenta, cyan, black and white was the usual combo.
Even so, I spent a massive amount of my childhood tinkering with it, from writing text adventures in GWBasic to playing games. My favourites were Targhan, Death Track, Speedball, Wizard Warz, and the old Sierra and LucasArts adventures. I also typed all my GCSE essays on the character-based Letter Perfect word processor.
Joe Martin, Games Editor
First PC: IBM P133
My first PC was an IBM P133; it was one of those beige desktop units that was laid on its side with the monitor positioned on top – a setup that you don’t see much any more. I don’t recall the other specs, and I can only remember the speed of the CPU because it was written on a sticker next to the CD drive. I fiddled with that sticker a lot while I swapped in and out the six CDs that held my favourite RPG, Betrayal in Antara. The P133 wasn’t the first machine to get me interested in computers though – that honour goes to the Amiga A500+ that preceded it. I may be the least technically competent member of the Custom PC staff, but even I was able to learn the basics of making my own simple games with the Amiga’s AMOS programming language. Later, I used PC software such as Klik ‘n’ Play to experiment further on the family IBM PC.
Paul Goodhead, Staff Writer
First PC: Anonymous 486-era PC
My first PC cost around £2,000 and was disappointing in almost every way. It was slow, loud, nothing ran on it and we had no Internet connection. My only fond memory is of playing Descent on it, though I was so young that all I did was blast away at anything that moved until I’d used up all my missiles.
It also needed occasional upgrading – an arduous and expensive task that involved leaving it at the local PC shop for a week. It would then come back plagued with compatibility and stability issues before being returned to the shop. This put me off PC gaming, pushing me towards consoles for the next few years. It wasn’t until I saw people gaming over a broadband connection at uni that I saw the potential for PC gaming and built my first rig.