30 years ago today, a new office automation device was officially launched by IBM at the Waldorf Astoria hotel in New York. August 12 1981 marked the launch of the IBM 5150 Personal Computer, a device which whilst not revolutionary itself, started a revolution in the way we manage and consume data.
Today Dr Mark Dean of IBM International Business Machines Corporation 152.49 -0.71 -0.46% thinks that the PC is now dead, and we should move onto something else, whilst Frank X. Shaw of Microsoft Microsoft Corporation 69.96 +0.34 +0.49% thinks we’re now in the PC-plus era, and not the post-PC era.
So how did we get here, and is the PC dead, or just transmuting into something else?
To my mind, the thing that made the PC great was that despite the fact that IBM could have made it something different, more powerful, more proprietary, they instead used the commodity Intel Intel Corporation 36.26 0.00 0.00% 8088 CPU running at 4.77Mhz (for comparison, my HTC Desire HD phone has a minimum clock speed of 249Mhz, and runs normally at 1Ghz, between 52x and 200x faster, even before we look at IPC (instructions per clock) and processor efficiency gains).
Mark Dean led the team that developed an expansion bus that was publicly documented. Of all the various elements in the PC design, I think the most important of this were the five expansion slots, using what would later become the ISA interface, which allowed the basic PC to be adapted to suit all sorts of different needs and requirements, and by opening the interface specification IBM allowed other companies to offer interfaces and expansion that IBM themselves hadn’t even thought of.
One of the slots would be filled with a Monochrome Display Adaptor (MDA), which also provided a parallel port for the connection to a printer (or a Color Graphics Adaptor (CGA) card running a prodigious 320×200 pixels in 16 colours, or 640×200 pixels in two colours), a floppy-disk controller supported Single Sided/Double Density 5 ?” floppy disk drives, at 160Kb each (later Double Sided / Double Density drives became available, increasing capacity to 320Kb, and with an extra sector to 360Kb). Serial ports would allow a connection to a modem for access to a number of services available at the time.
The other thing that made the IBM PC so great was the fact that both input and output were considered as part of the overall package. Output was handled by a bi-directional dot-matrix printer, and the input was via a good quality keyboard, drawing upon the heritage of IBM’s golf-ball “Selectric” electric typewriters. The keyboard had a good feel and quite a loud switch action, suitable for touch-typists. Indeed, I still use an ancient Compaq keyboard with a very similar feel, and according to my colleagues on conference calls, with an equally annoying click for each keypress!
Whilst Microsoft might claim that their operating system (MS-DOS) was the key to the PC’s success, I’m more convinced it was due to the availability of applications such as Visi-Calc, yet again proving that content is more important than the platform itself. MultiMate wordprocessor and dBase III were key applications for creating content, this was certainly the case when Lotus (now itself part of IBM) launched the Lotus-123 spreadsheet launched in 1983, which subsequently drove the demand for memory beyond the 640KB boundary available in MS-DOS (so we’re seeing within two years the demand for over 10x more memory than the basic PC started out with), and this later grew with the Expanded Memory systems and boards to page switch up to 4MB of RAM into spaces in the upper 384KB of memory that the original PC specification had held for expansion ROMs and I/O.
In March 1983, we also saw the launch of the IBM PC/XT (Model 5160) with a single floppy-disk drive and an internal Seagate ST-412 10MB hard disk with an Xebec hard disk controller using the ST-506 interface, and an increase from 64KB of RAM to 256KB of memory on the main board. This was in demand to the changes for extra memory (as above) as well as the fact that juggling floppy disks to boot a system, load an application, and then load a data file, work on it, and save it again, before moving onto a different task. By having all of these things installed on a hard disk made the process of using a PC for multiple tasks much easier. The expansion slots on this device has defined width and spacing of expansion slots on all subsequent devices, and is still used on motherboards today, even though the physical interfaces are different.
In 1984, IBM improved the PC, it became the PC/AT (Model 5170) with a new processor (Intel 80286 at 6Mhz), up to 16MB of memory, and the Enhanced Graphics Card (EGA) which provided 640×350 pixels with 16 colours. The extra memory was not directly addressable under MS-DOS, and had to be paged in using specific device drivers (HIMEM.SYS) and address-wrapping techniques.
But IBM wasn’t having everything their own way, as the choice of commodity hardware and open interface specifications quickly allowed copies to be made, but the one thing that wasn’t published was the BIOS specification. So this had to be engineered using clean-room techniques. Compaq (now part of HP HPQ 18.47 +0.10 +0.54%) quickly produced a BIOS that was very closely compatible with the IBM original, and it’s carry-top transportable form-factor (that the Osborne 1 pioneered a few years earlier) as the Compaq Portable. Olivetti produced the M24, which was an 8MHz Intel 8086 PC-compatible I’m intimately familiar with, since it was the first PC device I spent a lot of time with.
IBM however, made a mistake, in my mind, when they chose to become less compatible with other devices when they transitioned from the PC to the Personal System/2 (PS/2) environment, moving from the more open ISA (and EISA) buses to the proprietary Micro-channel Architecture (MCA) interfaces for expansion cards. Whilst this did allow them to gain more control in specific accounts, it lost them the technical advantage gained by the openness of the original PC platform, and as other options were available, customers moved to a more compatible PC-compatible, ironically no longer an IBM.
The revolutionary change brought by the PS/2 did bring some advancements that were rapidly integrated into the PC-compatible platform, including 31/2″ floppy disk drives, VGA (640×480) graphics and the smaller PS/2 6-pin mini-DIN keyboard and mouse interface, and the memory SIMM (rather than banks of SIPPs, or even individual RAM chips).
Other companies developed systems, motherboards, expansion cards, and even some companies even started creating the combined chipsets that took a lot of discrete components and joined them in a smaller number of integrated components. Amongst the first was the Chips and Technologies NEAT chipset, which allowed designers to deliver more reliable, cheaper motherboards. I know that it certainly simplified my life when I was working building systems at the time, with substantially fewer failures on first on install.
Indeed, integration has occurred on so many elements of the PC marketplace since then. For example, the three cables for a dual-hard disk ST-506 setup have been integrated into the single IDE PATA cable, and further into the single narrow SATA cable per drive. Integration has also move system sizes down from AT-sized boards to ATX to ITX sized platforms. The PC has now spawned a multitude of form-factors and capabilities, from desktop, to tower, to rack-mounted server, to laptop, to netbook, to home entertainment server, to gaming platform.
The other major integration that has gone on is the rise of networking and the Internet itself, integrating data and content across several separate devices, no matter where they are in the world, providing a way to share data and ideas between individuals. Without the ability to take the content and share it with a wider audience, the content would be consumed by a very small number of people, with the Internet and the World-Wide Web, content is created and shared globally and the PC was the first real platform to allow this to occur. Indeed networking is driving the next innovation, which is to embed data and content in the cloud, and the PC is just one of the surfaces that collects, edits and consumes content.
I don’t think that the PC is yet dead, but it has moved past being the single point of access to content, both recreational and within the enterprise, although it has helped blur the boundary between the two. The PC still has a future, but as part of a greater hegemony of devices for creation and consumption of content, with the data held within the cloud.
The PC lives, long live everything else!
Note: Just to prove a point, I compiled this post on a desktop, notebook and smart phone. Content is now everywhere!
Update: 16 July 2016 – changed stock references for HP to reflect Hewlett Packard Enterprise following the split of HP