Articles: Computers - A Bluffer's Guide Computers - A Bluffer's Guide The history of computing is a long, winding and, occasionally, very blurry tale. Just where do you begin? What actually constitutes a 'computer'? In short, how has mankind reached the point where our lives are now run by tiny slivers of silicon? Maybe you'll find the answers to these questions here (although probably not) but what follows is a rough guide to computing, the evolution of computing, and the major breakthroughs and milestones. Maybe we might just shed a little light on where your little Mac (or PC) came from. Note: Apologies in advance if I've missed anything/anyone but it's a big subject to cover, there's lots of events and technologies to cover and history does have a habit of becoming 'blurry' at times (who invented what and when is especially fun to unravel). As per usual all comments and suggestions are gratefully received and, hopefully, this won't prove to be too boring a read =;) The Dawn of Time So Where Did It All Begin? To work out when the computer began we have to define what precisely we mean by the term 'computer' - if we don't then we're going to spend an awful lot of time covering an awful lot of material that is of no interest to us. At the most basic level we could look at a computer as being a mechanism for counting - great, that leaves us looking at everything since mankind discovered that he had ten fingers and toes. Maybe this is a bit too far back so what's next? The obvious answer is the abacus but, again, this is still a bit too broad to be really practical to cover in any great detail. So where does this leave us? By 'computer' most people instantly picture circuit boards and keyboards. Occasionally someone will picture a room full of filing cabinet sized boxes and giant spools of tape, but this image is quickly dying out. In common language 'computer' actually means 'electronic, programmable computer' and that's really what we all want to know about, not fingers, toes and beads in frames. If we stick by this definition then we certainly have a framework in which to work and I don't have to start rambling about esoteric rubbish. So we have the 'electronic, programmable computer' but what does that mean and where did that come from? Mechanical Machines? Why, The Very Thought The concept of a creating a machine to perform calculations stretches back many centuries and the most prominent exponent of this (in the minds of most) would be Charles Babbage. Actually he wasn't the first and, all across Europe, mathematicians and invenventors dreamed of created machines that would perform calculations. German Wilhelm Schickard is generally regarded as the first person to develop a workable solution in 1623. Despite no surviving examples of his work, his notes and correspondence provided enough evidence and information for working models to be reproduced. This wasn't a problem that applied to Frenchman Blaise Pascal (1623-1662), a skilled mathematician who developed a mechanical machine to help his father calculate taxes. Developed in 1642, the Pascaline (as he named it) was a purely mechanical device that resembled the huge mechanical calculators of the 1940's but Pascal had managed to overcome the problems of the French monetary system (20 Sols in a Livre, 12 Denieres in a Sol etc.). There were high hopes for the Pascaline and Pascal produced 50 prototype machines for sale in 1642, but sales were very poor and the venture was doomed to failure, and the whole enterprise lasted less than a year. It didn't go unnoticed though and, in 1670, Gottfried von Leibniz took Pascal's ideas and refined them into his 'Step Reckoner' machine. Unlike the Pascaline which could only add and subtract, Leibniz's machine could multiply, divide and even evaluate square roots - an impressive achievement in an age when many people still thought that the world was flat. Skipping forward to the 19th century (nothing much happened in between) and Charles Babbage (1791-1871) revisited the idea of a mechanical machine that could perform calculations. At this time the industrial revolution was underway and mathematicians, navigators and every trade relying on numbers, was forced to rely on tables of calculations (e.g. Sin, Cosine, logs) - calculations that were slow, laborious to produce and relied on teams of people to calculate them (people called 'computors'). Acutely aware of the problem, Babbage (a self taught mathematician) set his mind to creating a machine that could do the job. The result was the Difference Engine in 1832. The Difference Engine was a major achievement but was limited in that it could only perform one mechanical task. A better solution would be a machine that could manipulate symbols in any manner and Babbage soon dreamed up The Analytical Engine - a project that he started working on in 1856. Babbage though was never able to complete his machine and this was not only because he died but also because he had a tendency to jump from idea to idea and never actually completed very much. That task fell to Swedish printer George Scheutz who did manage to build a machine based on Babbage's designs. Amazingly, despite the engineering limits of the time, the machine did actually work and produced mathematical, astronomical and actuarial tables which were far more accurate than anything seen before. After his death in 1871, Babbage's son continued with the Analytical Engine but it was never finished and, on the few occasions that it did run, it produced obvious errors - so much for steam powered British computers. Note: The British Science Museum set about the task of re-building Babbage's Analytical Engine (from scratch) in 1985 and produced a working machine in 1991. Ok it weighed 3 metric tons and had 4000 parts, but it did work. Punchcard To The Rescue The 'clever' readers out there will now assume that we'll jump forward into the 20th century but the steam powered 19th century still had one notable contribution to make. By the end of the 19th century, the US government had been taking a census every ten years and this was starting to become a problem. Originally a simple physical head counting process, the explosion in the number of people living in America had forced the government to re-examine the process as the amount of data being collected was starting to take nearly 10 years to process - by 1890 it was estimated that the results of the Census wouldn't be available until well past 1900. Enter Herman Hollerith and his automated electrical tabulating machine. The process was simple and, instead of counting heads or ticking boxes on a piece of paper, Hollerith proposed using an adaptation of Joseph-Marie Jacquard's punched card system (this had been used in French silk weaving since the early 1800s as a way of defining patterns in woven cloth). By having the data stored on punchcards, Hollerith detailed how the cards could then be collated and processed by an automated machine. By sorting the cards (by whatever criteria was needed) and incorporating a huge number of clock-like counters, Hollerith's machine allowed far more than just a simple population count to be performed and a far more detailed statistical breakdown of the American population to be compiled. The US government went for Hollerith's solution and were happy to finance the project. The results paid off, and, instead of taking years, the 1890 census results were processed in a matter of months. Hollerith had opened up a whole new avenue and statisticians quickly saw uses for the system. So much so that Hollerith setup the Tabulating Machine Company in 1896. In 1924 the company, having been amalgamated into by the Computing-Tabulating-Recording (CTR) group of companies, was rebranded as International Business Machines - IBM. With Hollerith's machines providing the processing power, the computing industry was happy but science wasn't able to sit still for very long. The electronic tabulating machines were a great invention but they were still very limited in what they could achieve and scientists and mathematicians around the world were starting to develop new and improved systems that pushed the limits of the fledgling computer industry. 1900 - A Brave New Century Into The Electronic Age...ish With the second world war pushing technology and science to the maximum, computing suddenly became a very hot topic and many, many breakthroughs and advances were made. So much so that history becomes very confused and clouded (especially with government security and the like). For years America claimed to have developed the world's first 'electronic programmable computer' in the form of ENIAC and the ABC, but Britain then assumed the title with Colossus. We'll cover all of these in greater detail later, but it now seems that a little German chap called Konrad Zuse may have beaten everyone to the title. Having been intrigued by the concept of an electronic brain, Zuse began constructing his Z1 machine in the living room of his parent's apartment in 1936. Having to rely on handmade metal plates (rather than the more usual electronic relays) the Z1 took two years to develop and was far from reliable. What it was though was inspiration and Zuse realised that by using relays instead of the handmade metal plates, he could construct a far more reliable machine. The resultant Z2 used 800 relays in its arithmetic unit and proved to be far more reliable and impressive. So much so that Zuse managed to receive a modest backing from the Nazi government. Beginning work in late 1939, Zuse constructed the Z3 and demonstrated it to an audience of scientists and mathematicians on May 12th 1941. Feeding the 'program' in, the machine burst into life and it's 2200 relays successfully produced the correct result. This was no fluke though and Zuse fed in several different programs, each of which resulted in the correct answer being produced on the four numeric lamps. The Z3 worked. Yes it was slow (it took about 3 seconds to perform each simple operation) but it did work. The Nazi government recognised this but failed to see the need for such a machine - the war would be won soon so what would it be used for? Eventually the Z3 did perform some calculations that helped provide wing stability in aircraft wing designs but, in a cruel twist of fate, the whole machine was destroyed by Allied bombing in 1944. Note: Just as with Babbage's Analytical Engine, the Z3 was reconstructed (by Zuse himself prior to his death in 1995) and now stands in the Munich Deutsches Museum. Although the Nazis failed to see the need or potential for computers, elsewhere the need was blatantly obvious. Before the second world war had even started, Britain had realised that it faced the problem of how to decrypt German messages and, in 1939, setup the Bletchley Park establishment who's job it was to solve the problem. The German government had relied on two encryption systems Lorentz and Enigma (later expanding into a second upgraded system nicknamed 'shark' by the German navy). The Polish Intelligence Service had been working on breaking Enigma since before the start of the war and had developed an electromagnetic machine that could repeatedly look for encryption keys - the Bomba. Having shared the Bomba with Britain, mathematicians Alan Turing and Gordon Welchman turned their minds to developing a better version. Although Bomba worked, it was not hugely efficient and often it could not provide an encryption key until after the key had been changed (the Germans wisely changed the encryption key daily), rendering any new messages undecypherable. Understanding how the Germans worked and how they used Enigma, Turing and Welchman created a more efficient Bomba (which they renamed 'bombes'). The bombes helped to solve the Enigma riddle but the Lorentz system was entirely different and even the improved bombes couldn't help with this. Used only by the most high ranking German officials (including Adolf Hitler himself) Lorentz appeared to be unbreakable. Assigned the task of breaking Lorentz, mathematician Max Newman managed to develop a system to crack the code, but it was time consuming and the resultant machine quickly got its paper tape mechanism out of sync. Realising that he had a problem, Newman contacted Tommy Flowers a Post Office telephone engineer, to see if he could come up with a solution. Flowers immediately suggested the use of electronic valves rather than the more unreliable relays that Newman had been using up to that point. With the pair working together, they managed to construct Colossus in 1943. By using valves, Colossus solved the synchronisation problem but was also far, far faster and worked around the clock from January 1944. Despite being designed to break the Lorentz code, Colossus was actually capable of being fed different programs (a feature that Turing's bombes totally lacked - they did just one job) and eventually ten such machines were built. Flowers and Newman had created what was considered, until very recently, to be the first 'electronic programmable computer' in the world...but the world would, sadly, never know about it. With the war won, British leader Winston Churchill ordered the destruction of the machines and everything related to Bletchley Park's work - a desperate attempt to maintain secrecy that worked for 30 years until Britain's rules regarding the declassification of material came into effect. Computing Grows Up Enter ENIAC Both the Z3 and Colossus had shown that electronic, programmable computers were possible but their achievements had not been revealed to the rest of the world. In 1943 the US government faced the same problems that had plagued Britain - how to get reliable calculations at high speed. With the war raging on, the US had developed newer and more powerful weapons and it was essential that battlefield users knew exact ranges and distances so that they could be used effectively. To date this job had been carried out by teams of female 'computors' but the volume of work was proving to be a major bottleneck. The answer was presented to them by John Mauchley and Presper Eckert who proposed ENIAC - Electronic Numerical Integrator And Computer. ENIAC would be unlike any (widely known) machine and would allow high speed processing and calculations that would be far more accurate and reliable than any human derived results. It wouldn't be cheap though (initially $61,700) and, after dragging on for three years and costing over half a million dollars (in 1940's money!) ENIAC was revealed to the world on February 15th 1946. Despite the fact that ENIAC had been developed without any knowledge of Colossus or the Z3, it still was not entirely 'unique' - even within America. Two machines managed to beat the 'first' that was ENIAC - the ABC and the Harvard Mark 1. Almost seven years before ENIAC, John Atanasoff and Clifford Berry had been experimenting with various concepts and, over the course of three years, had developed the Atanasoff-Berry-Computer - ABC. The ABC introduced many new concepts (parallel processing, separate memory and processing etc.) and, prior to the discovery of the Z3, was often thought of as the first electronic computer. The outbreak of the second world war though had greatly affected the machine's development and it sat in a basement in the University of Iowa, unpatented and unused. While the ABC had been more of a research project, the Harvard Mark 1 (officially known as the IBM Automatic Sequence Controlled Calculator) was a fully fledged digital computer. Unlike the ABC and most of the major machines that followed it, the Mark 1 relied on switches, relays, rotating shafts and clutches, instead of valves and more purely electronic components. The Mark 1 did work though and, although 50 feet long and 5 tonnes, was put to work by the US Navy until 1959. Developed by Howard Aiken and Grace Hopper, the Mark 1 differed not only in terms of technology, but also in terms of approach. Essentially many parallel calculators controlled by a single control unit, the Mark 1 broke problems down into smaller parts and then reconstructed them to form a final result. Different but still effective - and in place a year before ENIAC was unveiled to the world. Technologically it was a dead end though. ENIAC may have stolen the Mark 1's thunder (to a certain degree), but Aiken went on to make one of the greatest computer quotes of all time. Displaying an attitude which was typical of the time (i.e. very few understood the potential of computers) in 1947 Aiken confidently predicted that the world would need only 6 computers - 3 for America, 1 for Britain, and 2 for the rest of the world. How times would change. So why aren't the Mark 1, the ABC, Colossus and the Z3 immediately spring to mind when the 'first' is mentioned? The answer came down to, amazingly for a scientific endeavour, marketing. The Z3 and Colossus had been shrouded by wartime secrecy, and the same war had put the ABC into a basement. The Mark 1 didn't suffer from these to the same degree but it seems that only ENIAC really promoted itself to the world at large. But then again, ENIAC was far too big to remain a secret. Despite having many predecessors (although most were unknown to them) Eckert and Mauchley's ENIAC massively eclipsed all of them in terms of size and complexity. Its 80 foot length housed 19,000 valves, 1500 relays, miles of wiring and weighed in at over 30 tonnes! Yes it took 200 kilowatts to power it but running at 100KHz it was a quantum leap forward from the 5Hz speed that the Z3 boasted (a leap of 20,000 times faster). ENIAC wasn't all powerful though and, for all its valves and speed, it still lagged behind its earlier brothers. The Z3 had allowed programs to be fed in via a perforated film strip reader but ENIAC needed to be rewired via its enormous plugboard - hard and difficult work but ENIAC simply wasn't designed to store its programs in memory. Later enhancements would go some way to remedy this but in 1946 ENIAC was still hot property: While it took a skilled 'computor' 20 hours to calculate a trajectory, ENIAC could come up with the answer in 30 seconds (even if the actual flight time for the projectile was just 15 seconds)! Interest in ENIAC grew and several projects sprang up around the world almost simultaneously (and, more often than not, blissfully ignorant of each other). From Russia's highly secretive machines (used for atomic bomb calculations) to the groundbreaking work done in Australia, computer science was acting in an almost digital way - one minute it didn't exist at all, the next almost every country was developing a machine of some kind. Of course most adopted the 'AC' naming convention (with one of Australia's machines being named, rather unfortunately, 'SILIAC'). One of the most remarkable and recognisable machines that emerged after ENIAC was Manchester University's Small-Scale Experimental Machine (SSEM) more affectionately known as 'Baby'. Although Baby was tiny when compared with ENIAC (both physically and in terms of processing power) it did provide a testing ground for many of the technologies and techniques that are still with us today - many of which had been proposed but not yet implemented. Unlike ENIAC, Baby could store both data and a program, and it also boasted 'true' Random Access Memory (RAM). These were revolutionary concepts in 1948 but they proved that it could be done. And if it could be done on a small scale then it could be done on a big scale (as ENIAC had proved). The Manchester 1 was born offering far more power and usability than Baby could offer. So impressed were the British government that they commissioned Ferranti to build a commercial version of the Manchester 1 - and so the Ferranti Mark 1 was born. Technically the Ferranti was the same as the Manchester 1 but a couple of tweaks here and there upped performance, and made it the world's first commercially available computer (with the first one being delivered to Manchester University to replace...the Manchester 1). While ENIAC may not have been the first, it was certainly the most well known machine in the world of 1946, and it raised the profile of computing enormously. Even in it's very early days (the machine would eventually run until late 1955) ENIAC garnered much interest in the scientific community (spawning the likes of Baby) but it wasn't just white coated boffins who became interested. The tea shop owning Lyons company of Great Britain had wrestled with how to streamline and improve their business for many years and, in 1947, it seemed that America might offer a solution... Into The Mass Market The Lyons company needed to manage not only its vast string of high street tea shops, but also the production, transport, ordering and distribution of all of their products (which were manufactured in-house rather than bought from suppliers). By the 1940's this was becoming a gargantuan task that was prone to error. Despite being an old, established, family run firm, Lyons bought into the idea of having a machine that would not only speed up the process but also improve accuracy, and the fledgling team setup by Lyons made several trips to America to visit the likes of Eckert and Mauchley. Unaware of just how close to the cutting edge they were, Lyons eventually built the Lyons Electronic Office - LEO. Working with Maurice Wilkes from Cambridge university's newly formed computing department, Lyon's co-developed and co-financed EDSAC - a machine that would eventually evolve into LEO. EDSAC took concepts and ideas that had been developed following ENIAC but so far they had remained as nothing more than ideas. The combined team turned concept into reality and, in May 1949, EDSAC ran its first calculation. Completed in 1951, LEO may have been beaten to market by Ferranti but it performed (arguably) the first regular business job undertaken by a computer. Yes ENIAC had been commissioned by several organisations to perform calculations but LEO was dedicated to the task of running business calculations. Being easily programmable, LEO offered a flexibility that allowed it to take over more and more tasks within the Lyons organisation but then Lyons made a jump. Formed in 1954, the LEO Computer Company became an offshoot of Lyons and started to supply machines to other companies. The original LEO was an impressive machine boasting 5936 valves, swappable banks of components, and 64 mercury delay tubes (used for storage) - quite an achievement when you consider that each tube was 5 feet long and weighed half a tonne! In addition to this, LEO used several oscilloscopes and even a speaker so that operators could monitor (and 'listen') to what was going on inside the machine. Over time some of the programmers became so accustomed to the sounds that they used the speaker to generate some of the very first 'computer music' - not quite a rival to Apple's iTunes, but a start. With so many gains being made in so many areas, keeping track of what was going on was becoming difficult (believe me, writing and researching this is a nightmare) but the computer field was accelerating rapidly. Britain was producing machines for the commercial market but America hadn't simply sat back with the completion of ENIAC. The VAC Age Having created ENIAC for the US government, Eckert and Mauchley soon decided to push forward with new designs and machines. Sadly their military backing quickly started to dry up following the end of world war 2, and the pair decided to open their own business designing, manufacturing and selling machines. ENIAC had been a wonder but the pair knew that it was not the way forward - besides which the new EDVAC (Electronic Discrete Variable Automatic Computer) would be a far more powerful and useful machine. EDVAC addressed many of the problems that ENIAC had suffered but it would not be an easy machine to produce. Having lost several key team members, progress of EDVAC was slow and, although conceptually designed in 1946, the final machine wasn't available until 1952. EDVAC's main advance over ENIAC was the used of stored programs. ENIAC had relied on plugboards and the operators having to (literally) hard wire in the program. This was slow, inefficient, restrictive and very time consuming (it could take weeks to rewire the machine) and the concept of 'stored programs' had been developed by Eckert, Mauchley and Professor John von Neumann. Instead of hard wired programs, EDVAC would store its programs in memory, meaning that to reprogram the machine you simply had to change what it was storing - far quicker and easier. The machine would then simply perform it's current task and then ask the memory what it was supposed to do next, perform that operation and then ask what it was supposed to next...more jazzily known as the 'fetch-execute cycle'. Documented by von Neumann (who now appears to have neglected to mention the contribution of anyone else) this approach would become the underlying concept of just about every machine in the world, even being employed in modern desktops (and is still referred to as 'von Neumann architecture'). The concepts were there for all to see and EDVAC would be the first machine to make it work...sadly there was competition. Manchester's 'Baby' had proven that it could be done, and Cambridge and Lyons had built the EDSAC on the same ideas. All of them beat EDVAC in the race and Lyons' LEO would even be running business software a year before EDVAC was even completed. Worse still was that EDVAC was beaten by it's own offspring - UNIVAC 1. Having hit financial problems, Eckert and Mauchley's 'Computer Corporation' had been acquired by Remington-Rand and the big corporation wanted something that it could get out onto the market (as opposed to a research project). Based on the EDVAC design, UNIVAC was a fully operational, commercially available machine that was, amazingly, available a year before it's 'big brother' was finished (the first machine, UNIVAC 1, was installed at the Bureau of Census in 1951, and boasted the use of a magnetic tape unit as a memory buffer - another first). Eventually completed in 1952, EDVAC did not go on to a distinguished career and a 1956 report showed that it achieved an average error free up-time (the time between errors/crashes/failures) of only 8 hours. There's No Business Like Showbusiness UNIVAC may not have been the first stored program machine but, like it's 'grandfather' (ENIAC), it quickly gained a far better profile and garnered more media attention than anything else in the world. The 1950's were the age of television and UNIVAC was about to be demonstrated to the biggest captive audience on the planet. Public interest and information about the infant computing industry was limited to say the least, and it was only ENIAC that had managed to make the leap from the laboratory to suburbia (in terms of information - at 30 tonnes, physically it wasn't going anywhere). Yes there were many developments taking place around the world (including several in America) but public perception of the power of computing was extremely limited. The computer was a great behemoth that was tended to by armies of white coated technicians; an electronic brain that would solve the mysteries and problems of the world; a scientific marvel only understood, and used by, the greatest minds. What it wasn't, was something that the man in the street had anything to do with or use for. The technology was so new that few understood the full implications or potential possibilities (hence Howard Aiken's estimate for how many machine's the world would need) and fact quickly blurred with imagination, speculation and outright fantasy. The computer really would do everything that was promised of it and, in 1952, fantasy seemed to become reality...and all in front of a waiting television audience. The 1952 race for the White House was destined to be a close run affair and Remington Rand approached CBS with the claim that their new UNIVAC machine could accurately predict the result before the votes had been counted (using previous voting patterns etc.). Expectations within CBS were low but they decided, that for sheer entertainment and curiosity value alone, it might be an audience winner. Moving UNIVAC (only a single version existed at the time) would be a very difficult and dangerous operation so instead the machine stayed in its Philadelphia home while a fake control panel was used in the studio. So, come election night, UNIVAC was fed the data and reported that Eisenhower would win by a landslide. CBS were doubtful and even refused to report it - the machine simply must be wrong. Time though would soon prove that man was no match for machine. With results coming in, CBS realised that UNIVAC was actually looking to be a lot more precise than they'd first though - even to the point where the studio confessed (live on air) that the machine had predicted the result hours earlier. The audience had just had their first glimpse at the power of computing and, for many, this was all that they knew about this strange new world of 'electronic brains'. With UNIVAC having proven itself in 'public', the general perception of computing changed almost overnight. What had been almost unknown to most people suddenly became fascinating. So much so that most people actually thought that UNIVAC was the computer - there simply was no other machine (when IBM released it's 701 a few months later, many people referred to it as "IBM's UNIVAC" - a term that certainly didn't please the tabulating machine giant). If UNIVAC could predict an election result, what else could it do? Suddenly computers could do anything and all manner of bizarre and unlikely tasks were thrown at the monolithic machines. With hindsight some are more fanciful than others but, as with any new science, computing was filled with unknowns and no limits. Having taken the concept of the 'electronic brain' onboard (along with the term 'artificial intelligence' coined in 1956 by John McCarthy) it wasn't just the general public who were baffled and misled by what was possible. The US Army set to work developing a system that would automatically identify tanks and field artillery in reconnaissance photos (using a very rudimentary form of artificial intelligence and neural-network coding). Amazingly the system did work and the machine was able to distinguish photos with tanks in them and photos without tanks in them...until someone fed in a new set of photos which it spectacularly got wrong. The original photos with tanks in them had all been taken on sunny days while the photos without tanks had all been taken on cloudy days. The machine had simply 'learned' to identify sunny and cloudy days. The project quickly disappeared but perfectly demonstrated the gap between reality and perception. ENIAC, the Mark 1 and many of the early machines, had all been designed for one customer to perform a limited number of tasks (sometimes just one task). UNIVAC, on the other hand, joined LEO in being a product that any company could buy and use for whatever purpose they could think of (eventually a total of 46 UNIVACs were sold). This concept of the computer as a product was certainly new but the industry was about to get its first major player - IBM. Note: Following the introduction of the UNIVAC and the entry of IBM into the marketplace, the number of computers and manufacturers increased to near stupid levels. Although so far we've been covering (hopefully) everything, from here onwards we'll have to stick to just the 'important' machines and events (although possibly not in any sensible order). More Than Just A Research Tool Enter Big Blue Herman Hollerith's automated tabulating machine (that had come to the rescue of the US Census Bureau in 1890) had led to the creation of IBM, and the company had made a steady profit out of the (now) ageing tabulating machine. Despite legal problems in the 1920's, the company had prospered under the rule of Thomas J. Watson Sr. and had even dabbled in the field of digital computing (producing the Harvard Mark 1). By and large though, IBM had no great interest in the computing field but, for a business that was based on paper based offices, the new theories that expounded the 'paperless office' were a big worry - who would need IBM's paper based products when offices didn't need paper? This was a concern but the outbreak of the Korean conflict in 1950 brought the issue to the forefront. Keen to help the war effort, IBM approached the US government offering their services. The response: build a large scientific computer to help aircraft design, atomic bomb calculations etc. While the company had constructed the Harvard Mark 1 and IBM's own Selective Sequence Electronic Calculator (SSEC), creating this new machine would be something different - it would need to be capable of being mass produced. After visiting various defence and aircraft manufacturers, IBM realised that, given the number of orders they'd already received (for a machine that didn't even exist), they had a lot of catching up to do. By early 1951 the new project needed to be started as soon as possible and IBM decided on a very rapid development and production plan - so rapid that development would take place while the machines were actually being produced. Design started in February 1951, construction (of the prototype) began just a month later, and the first production machine started being built in June. By December, the very first machine was completed and IBM labeled it the 'IBM 701'. Installed at the company's world headquarters in New York, the machine replaced the existing SSEC (literally - the new machine was placed in the same spot as the old one). Despite being just one quarter of the size of the old machine, the new 701 ran 25 times faster. Eventually the installation wasn't completed until 27th March 1953 (by which time UNIVAC had stolen IBM's thunder) but the 701 marked a new era in the computer industry. Technically the 701 incorporated many advances, but the actual involvement of IBM was vitally important too. Although Remington-Rand was doing a good job of selling and promoting UNIVAC, IBM suddenly created a presence that brought the computer into the office (rather than the research facility). IBM's network of salesmen immediately enabled the company to understand what customers wanted and, actually make customers a) aware of what the computer could do and b) that computers were no longer the stuff of science fiction. We won't really go into this aspect in any great detail though. From a technical perspective the 701 introduced the world to many concepts that had, to date, been little more than ideas and research projects - magnetic tape (housed in filing cabinet sized boxes of spinning reels) and cathode-ray tube storage (an idea originally employed on Manchester's 'Baby') being the most obvious. All of this technology would have been useless though if the 701 couldn't actually compute...but it could. Running at over 16,000 additions/subtractions a second, the 701 was the start of something very important. UNIVAC may have put computers into the public eye, but the 701 made sure that its dominance would be brief, a fact confirmed when the 1956 presidential election took place - IBM computers predicted the results this time. The 701 had brought IBM to the market but the new market leader wasn't about to rest on its laurels and the 701 actually only enjoyed a very short production run with only 19 actual machines being built (for cost reasons IBM also offered the option of 'renting' - at a mere $16,000 per month) and most of these ended up in government departments and the really big corporations. It's replacement, the 704, was released onto the market in 1954 and enjoyed 6 years of life, but it brought more than just a name change. More Power...To More People The IBM 701 had relied on a cathode-ray tube memory storage system and, while it worked, the system was not ideal. The 704 replaced this with a new type of memory - the high speed magnetic core (originally developed for the Whirlwind machine in 1951 but never used 'commercially'). By stringing thousands of wires together and combining electrical pulses, the system could store binary values. The big advantage was speed and, using a magnetic core, a number could be retrieved in just 12 millionths of a second - a significant increase over the existing methods available. The 704 was therefore much quicker when it came to its memory but the machine also introduced large scale floating point operations - mind numbingly boring but actually very useful. The market leader wasn't just concerned with hardware though and in 1954 the FORTRAN (FORmula TRANslation) programming language was created by John Backus. Before FORTRAN, programming had been a very laborious and difficult task. The likes of ENIAC had required programs to be entered in pure binary (a stream of 1's and 0's) and, with the current state of machines, tracing errors could take days or even weeks. EDSAC's Maurice Wilkes had pushed binary (first generation) programming forward with the concept of 'assembler' (second generation) and this helped enormously. Ok so assembler wasn't perfect but programmers could now enter more meaningful statements and these would then get translated into pure 1's and 0's. Yes you still had to understand how memory worked, you still had to move chunks of data from one place to another but it was, by comparison, a doddle. FORTRAN expanded on this even further. A third generation language, FORTRAN created a bigger cushion between the programmer and the hardware and there didn't need to be any (or, at most, very little) understanding about what the hardware was doing - you specified what the machine had to do and then this would get translated into the final 1's and 0's. FORTRAN was certainly one of the first third generation programming languages but it certainly wasn't the last. LISP (also developed by IBM) and ALGOL soon appeared, but it was in 1959 that the first really widely embraced language appeared: COBOL (COmmon Business Orientated Language). As it name suggested, COBOL was very much aimed at business orientated software problems and, compared to binary and assembler, was a revelation. So much so that the language, although now defunct (45 years later) still appears on many systems and provides plenty of headaches for programmers around the world. These third generation languages were starting to open up computing, and programs that would have been impossible to achieve previously, suddenly became realistic. They also allowed the theoretical possibility that a program written for one machine could be run on another machine - this was still the stuff of dreams though. Note: There were a wealth of computer languages developed but only a few still remain and are in widespread use. FORTRAN survived into the 1970's but soon after became a far more academic tool. LISP remained a purely nightmare collection of brackets and insanity (says one who speaks from experience). COBOL lumbers on and, like smallpox, occasionally rears its ugly, longwinded head. BASIC (Beginners All-purpose Symbolic Instruction Code) introduced generation after generation to programming and really gained exposure in the age of the microcomputer. C was developed and would become the most widespread (and abused) language in the world. And finally PASCAL (developed in 1969 and named after our old friend Blaise Pascal) became one of the best programming languages for beginners ever as well as becoming a very solid and reliable 'real world' programming language (it's also pretty handy at paying the bills - I'm a Delphi/Pascal programmer by day). There were also a wealth of others but I really don't have the time, enthusiasm or patience to document them all - apologies if I've missed your 'favourite' out. IBM were now firmly in the driving seat in just about every aspect of computing, and the newly named Sperry-Rand (following a merger between Remington-Rand and Sperry Gyroscope) soon found that UNIVAC was slipping behind the times. Yes UNIVAC had been the first machine in the world to have a high speed printer, but the change of pace was quickly leaving the once synonymous name firmly lagging behind. Dominant was the word that IBM had desired for a long time and by the end of the 1950's, dominant is what they were. The introduction of the IBM 1401 in 1959 simply re-enforced this as over 10,000 of them would eventually be sold during its lifetime. Machines were getting bigger and faster but they still had a habit of only doing one job at a time - yes they could do it faster and more accurately, but they still did just one job. In 1961 IBM's 709 and 7090 machines managed to change this...with a little help from MIT. Having developed a 'timeshare' system, several users could seemingly use the machine at the same time. While this wasn't exactly crucial in research establishments and business, in less high-pressure environments allowing only one user to 'use' the computer was very costly (this machine had cost a lot of money so you really did want to use it to its maximum). Timesharing worked by allocating small slices of the computer's time to each user so that they all had a little bit of processing done (the same principle is still employed in the likes of Windows only now the computer's time is split between applications rather than users). User A would start his program at the same time as users B and C, and the computer would work on program A for one chunk of time, then spend one chunk of time on program B, and a third chunk on program C. The cycle would then get repeated until everything was finished. This wasn't brilliant (as there was a lot of data being moved about as the computer switched from task to task) but by keeping the slices of time very small (as in hundredths of a second), each user felt as if they were had the total power of the computer at their disposal. This not only made more economic sense but it got more users working and experiencing computing - the appeal and widespread contact with computing was starting to spread. Welcome To The Option Age While the competition started to feel the heat and struggled to keep up, IBM were pouring vast amounts of money into not only their current products, but also future products. All machines to date had come in one flavour - you bought a 701, you got a 701, you bought a UNIVAC, you got a UNIVAC. This was fine for businesses and organisations that had sufficient money but IBM realised that there were other markets that simply couldn't afford the millions of dollars that it cost for a computer. The cost was justifiable as every machine was a self contained system that performed calculations as quickly as it could. Also, each model was entirely incompatible with every other model so if you needed software writing, you had to create it from scratch - there simply wasn't any concept of swapping code between machines. The solution was to stop thinking of each model as a standalone, and instead create a 'family' of machines. Announced in 1964, the IBM 360 series was available in a range of configurations. While the 360/20 was the lowest specced machine (and therefore cheapest), it still had the same instruction set as the top-of-the-line 360/91 so they could both (theoretically) run the same software. The difference lay in what was (and what wasn't) under the hood. In the case of the 360/20, as much processing as possible was done via software. This cut costs but also reduced the machine's speed. As customer's went further up the product line, more and more of the functionality was handled by hardware (as opposed to software) - faster processing but greater hardware cost. The project was risky though and, at $5bn, failure would have wiped IBM out. The gamble paid off though and, with it's first 360 machine (the 360/40 model) shipping in 1965, IBM had, once again, changed how computing was viewed and actually paved the way for the way we view and use computers to this very day. Unlike the machines that preceded it, the 360 needed to employ an interface and environment that was common across a range of machines. ENIAC, UNIVAC and even the 701, had all been very much bespoke creations, and understanding one had been of little use when users switched to a different machine. The 360 was designed to avoid this and users could switch between different models without the need to retrain. The key to this was the use of the common instruction set but to access this IBM created three different 'operating systems' (this was the theory but in practice the lowest specced 360 models tended to run specialised applications rather than the more general operating systems). By offering this range of machines that were actually compatible with each other, IBM opened up computing to a far wider audience. Companies that couldn't afford the exorbitant fees demanded for a UNIVAC or a 701/704 were now able to get on the bandwagon. Admittedly the lower end machines may not have been in the same league as the most costly machines but they did do the job. There was also the benefit that, as time passed and technology improved, the machine could be replaced with a more powerful one but without the need to retrain users, rewrite all of the existing software, or replace all of the other devices that went with it - it could simply be replaced with a new machine! So successful was the IBM 360 that other manufacturers quickly adopted the approach used and the age of 'standalone' machines was well and truly gone. I Can See Clearly Now For all its many successes and advances, computers were still not the simplest things in the world to use. ENIAC's giant plugboards had been long since replaced but computers still consisted of hundreds (sometimes thousands) of switches and lights. Punch cards could be fed in at one end and punch cards could come out at the other end, but computing was hardly a hands-on experience. In 1951 MIT's Whirlwind machine looked at a new way of man and machine working together. Designed by Jay Forrester, Whirlwind used several revolutionary ideas (magnetic core memory being just one of them) but its biggest contribution was its interface. Using a round CRT (Cathode Ray Tube) Whirlwind allowed users to view and interact with data rather than have to interpret it from a row of blinking lights. Whirlwind was intended for use with aircraft and, via the screen, users could use the attached light pen and select items (all of which were displayed). This was big news as the computer had to react to the users input rather than just work through a rigid program. The light pen was ideal for this kind of work but keyboards would become the de rigeur input device...even if it wasn't UNIVAC or IBM that adopted either it or the CRT. Founded in 1957 by Ken Olson, Digital Equipment Corporation (DEC) knew that it couldn't compete with IBM when it came to the massive mainframe machines and instead developed a new class of computer: the minicomputer. Released in 1960, DEC's PDP-1 (Personal Data Processor) was strictly small scale compared to the vast room sized mainframes but it provided plenty of processing power for a fraction of the cost. This cost wasn't only in terms of the hardware itself but also in environment - whereas mainframes required air-conditioning, modified power supplies etc., the minicomputer didn't and would quite happily sit in the corner of an office. The big selling point (according to DEC) was the PDP-1's expandability. Offered with a vast array of input and output devices, the PDP-1 shipped (as standard) with an alphanumeric typewriter, but it was the optional CRT displays (there were several different types available) that really set the PDP-1 apart from its predecessors. Oscilloscopes had been used for feedback for many years and in many machines, but the PDP-1's CRT was far more than just a monitoring device and followed the path set by Whirlwind. Just as with Whirlwind, users interacted with the machine rather than wrote a program, loaded it, sat back while it ran, and then analysed the results. The difference though was that this was a commercially available machine rather than a research/government machine. In terms of performance, minicomputers were no match for mainframes but their size and price made them very attractive to potential customers who couldn't afford the might of the IBM 701, 704, UNIVAC etc. Universities and other cash strapped organisations embraced the little machines and the DEC PDP soon became the source for the legendary Space Wars computer game (admittedly not the first computer game but certainly one that's gained widespread visibility and stood the test of time). The PDP-1 proved to be a very successful commercial success and DEC released a stream of PDP machines right through into the 1970s. While many people had initially confused 'UNIVAC' with 'computer', DEC's PDP machines opened up the market and brought computing just that little bit closer to the ordinary man in the street. Yes it was mainly to students and academic types (ordinary academic types as opposed to just high ranking Nobel prize winning type academics) but, of course, the students of today are the industry leaders of tomorrow... The Big Machines Start To Get Small Chips With Everything IBM's 360 'family' was a great innovation from a commercial perspective (in that it offered flexibility and acknowledged and addressed the minicomputer) but it also adopted and used the very latest in technology - integrated circuits (IC). Just as valves (as used in the likes of ENIAC) had given way to transistors, so too did transistors give way to the ICs, more commonly known as silicon chips. Valves (aka vacuum tubes) had been the mainstay of the electronics industry ever since the start of the 20th century, but in 1949, William Shockley of Bell Telephone Laboratories succeeded in creating the world's first transistor. Valves may have worked but they had many, many problems: they were large, unreliable, power hungry, and produced an awful lot heat. Transistors, by comparison, were miniscule, long lasting, produced less heat and consumed less power. The potential of the transistor was quickly realised and the behemoths that were ENIAC, EDSAC and co. were soon being replaced with '2nd generation' transistor based machines that were smaller, more reliable and faster - even IBM spotted the potential and released their first transistor based machines, the 1620 and 1790 (even if it wasn't until 1959, a full year after Seymour Cray built the first fully transistorised computer for Control Data Corporation - the CDC 1604, which became commercially available in 1960). Not that this was the only application for the new transistorised technology and anything that had needed valves (TV, radio etc.) was soon replaced with a version that didn't. Transistors looked to rule the world but they weren't without their problems. Just as with valve technology, transistors still had to be physically connected to each other to build the necessary circuitry and this, more often than not, involved engineers manually soldering components together. Not only an arduous task (given the complexity of the machines at the time) but also one where a bad solder joint could quickly render a circuit useless (or, worse still, looking fine but capable of producing gibberish). The challenge was to come up with a solution that eliminated the need for soldering and therefore removed the potential for broken connections. Texas Instruments (TI) was working on a solution supported by the US Army. Called the 'Micro-Module', the intention was to create a standard size and shape for each component and incorporate the necessary wiring. When completed, the components could then just be snapped together without the need to manually wire them together. The idea was a sound one and TI was making progress but new employee Jack Kilby thought otherwise and came up with the concept of creating a ready built package based on semi-conductor technology (which, at the time, was in its infancy). Kilby may have come up with the idea but it wasn't new, having originally been conceived by British radar scientist Geoffrey Dummer. Dummer failed to make a working system but, in 1958, Kilby constructed and demonstrated the world's first semi-conductor. The technology was crude and bulky but it did work and a patent was granted in 1959. Like so many technologies, understanding of its potential was practically non-existent and, aside from the US Air Force who showed some interest, the computer industry reacted skeptically. Not that Kilby was alone and, in 1961, Robert Noyce also filed a patent based on semi-conductor technology (although Noyce's later patent was for a more complex 'unitary circuit'). Despite the fact that the pair had not worked together, Noyce had come up with the same idea and, thankfully, a new breed of customer had emerged. The computer industry had been cool to the uses of the silicon chips but the aerospace industry (especially NASA and its upcoming Apollo programme) had great use for computing but the physical size of the available solutions made it impractical - the tiny size and power needs offered by silicon chips answered those problems. And, of course, once the potential was demonstrated, it didn't take long for IBM and co. to come knocking. Silicon chips offered enormous benefits to computer construction and once building sized machines could be shrunk enormously, the power demands were slashed, production became cheaper (and more efficient) and the end products were more reliable and affordable - exactly the things that IBM needed for the 360. Where Next? Homebrew Anyone? The IBM 360 had not only introduced the concept of a distinction between the hardware and the software (in that the same software could be made available on a range of different hardware) but it had also spawned several other computer companies, and created the concept of different sized machines (e.g. 'mainframes' and 'mini computers' - mainframes being the big workhorse machines in the same style as ENIAC, minicomputers being smaller machines). It was still the market leader though and, in the summer of 1970, announced the a replacement to the 360 - the 370. The 370 improved on what the 360 had achieved, adding new instructions and hardware options (most notably a floppy disk drive which actually loaded the 370's microcode at startup), and allowing concepts such as virtual memory (in that the machine could treat offline storage (such as magnetic tape) as 'extra' memory above and beyond the memory that was physically available). What still didn't exist though was compatibility between manufacturers but IBM's ex-chief 360 architect Gene Amdahl (who'd also developed the first operating system for the IBM 704 way back in 1954) had ideas in that direction. Having left IBM in 1970, Amdahl started creating direct clone machines based on the 360 and 370 series, eventually shipping the first Amdahl machine in 1975. This was a major breakthrough and, for the first time, software from one manufacturer could be used on hardware from another. Hitachi quickly followed suit and soon released their own 'IBM compatible' machines. Once again, Big Blue had achieved another iconic feat...even if it was inadvertently. By the 1970's, computers had shifted out of the laboratories and international corporations, and were now appearing in mid, and even small, scale businesses. White laboratory coats were out and polyester clad office workers around the world could enjoy the benefits of the 'paperless office' (which seemed to produce more paper than ever before). The massive benefits and potential of the silicon chip had revolutionised the industry but it still hadn't filtered down to every level of society. Although office workers could harness the power of the digital age, the man in the street was still outside of the digital revolution. The integrated circuit had brought down costs enormously and some complete low end processors were appearing and becoming affordable (dropping down into the hundreds (and, sometimes, even tens) of dollars). Robert Noyce had founded Intel in 1968 (with Gordon Moore) with the intention of producing memory chips but dealings with Japan had resulted in the creation of the world's first complete processor on a single chip - the microprocessor. The fledgling Intel had been approached by Japanese company Busicom to produce 12 custom chips that would form the heart of a new electronic calculator that the company wanted to produce. Intel engineer Ted Hoff thought differently though and reckoned that a single general purpose chip (programmed via software) could do the job just as well. Busicom were convinced and the two companies worked on the plan for nine months, eventually giving birth to the 4004 microprocessor - the Busicom 4004 microprocessor (Intel eventually had to pay up $60,000 to buy back the design and marketing rights). The Intel 4004 was a revolution, cheap to produce and very small, but, as it could only handle 4 bit values (i.e. values from 0 to 15), it was kind of limited. The 8008, released in 1971, addressed this issue and allowed 8 bit values to be used (i.e. values from 0 to 255) - far more useful and enough processing power to allow it to do something 'useful'...like become a complete computer. Note: In 1965 Intel's Gordon Moore predicted, that chip complexity (i.e. the number of transistors incorporated into a single chip) would double ever couple of years - a prediction that still holds true over 40 years later. The microprocessor brought real power to the industry but, as per usual, the industry didn't really seem that bothered. With their giant mainframe and mini-computers, the power of the microprocessor was miniscule by comparison so what was the point. The computer manufacturers had, once again, failed to see the potential - "Who on Earth would want to what use that?" While home users could never hope to achieve the levels of computer use/reliance that had been promised for years by the media (fully automated homes, computer controlled lives etc.) groups of enthusiasts, fuelled by exposure to computers through college and work, were starting to emerge who wanted to harness the power of computing for themselves. Built in garages using the most basic of materials, machines were starting to appear that could do things. Yes they were extremely limited and lacked displays (usually relying on flashing LEDs) and input (a couple of toggle switches instead of a keyboard) but they did work. Electronics magazines started to include features on computers and, in the 1974 July issue of Radio Electronics, an article showed how you could build your very own computer (based on the Intel 8008, despite the fact that Intel had released its 8080 - cost was still an issue though as the 8080 cost $120). They called it "your personal minicomputer". Home users wanting a computer had few options in 1974 and, despite the French Micral machine (which had practically no impact in the all important US market) and the Scelbi 8-H (which faltered due to limited distribution) the only real option was to design and build your own. In 1975 though the MITS (Micro Instrumentation Telemetry Systems) Altair offered a new solution. Considered by many to be the first microcomputer, the Altair was available in either kit form or ready assembled (for a cost of $495 compared to $395 for an unassembled one) and happy customers received either a ready to go machine or a big bag of bits and an instruction guide. It wasn't fast, it wasn't brilliant, but it worked and that was the most important factor - within three months 4,000 customers would happily agree. Using only LEDs for output, toggle switches for input (the machine had to be programmed in binary) and initially shipping with a mere 256 bytes of memory, the Altair was basic (to say the least) but its design did incorporate several expansion slots and this quickly led to peripherals and new features/options cropping up (which were kind of useful once the 'joy' of coding simple programs in pure binary wore off). Where one went though, others followed and MITS was quickly joined by IMSAI and a whole host of smaller companies operating out of garages and spare rooms. In one spare room a fellow by the name of Steve Wozniak was toying with the idea of building his own machine. Despite wanting to use the more powerful Motorola 6800 processor cost forced Wozniak into using the cheaper MOS 6502 ($25 versus $175) even though he'd already designed his computer on paper using the more expensive option. The result was the Apple 1 and teaming up with his friend Steve Jobs, the machine was sold to Paul Terrell, owner of The Byte Shop - the first home computer store. At this time Wozniak worked for Hewlett Packard and the pair had already approached the hardware manufacturer, looking to sell the Apple 1. Once again big business had failed to grasp the concept and had decided to pass, forcing the two Steve's to finance the Apple 1 themselves. Having ordered 50 Apple 1's, Terrell quickly sold out (despite the fact that the 'assembled' machines were little more than circuit boards - not even a case). Wozniak knew that the Apple 1 was a solid little machine but it wasn't everything that he hoped for and started working on the Apple II. Eventually nearly 200 Apple 1's were sold but the Apple II promised to be a totally different phenomenon. Assembled in a plastic case, complete with keyboard and power supply, the Apple II used a TV for it's display and could load and save data to magnetic tape. This was nothing that the Apple 1 couldn't do (once Wozniak had added a tape interface) but this time Apple actually supplied everything - Apple 1 buyers had to add the extra bits themselves. Coming with colour, sound, more memory and running faster, the Apple II was a major step up from the Apple 1 (and certainly a world away from the likes of the Altair) and users could more easily program and use the professional looking machine. In 1977 Apple was a small firm just like so many others, but in the Apple II it had a very saleable product and the paying public were more than happy to buy it (especially when Dan Bricklin and Bob Frankston released the world's first spreadsheet application, VisiCalc - only on the Apple II initially). The home user industry had now been born and the basic collections of hand soldered components had evolved. The very first machines (often housed in wooden boxes) had now become professional looking consumer units that appealed to more than just bearded electronics enthusiasts. Users no longer needed to be adept with a soldering iron or understand the intricacies of microcode, binary shift registers, or the principles of von Neumann architecture. Instead the computer was something that you could simply switch on, load up your favourite program, use, and then switch off. The age of the microcomputer and the home user had well and truly arrived. Computing For Everyone The Empire Strikes Back IBM's 360 and 370 series of machines, along with their various clones, imitators and competitors, had been somewhat cool towards the emerging microprocessors being developed and released by Intel, Motorola and the like. These tiny slivers of silicon were no match for the filing cabinet sized machines and there was no perceived value in developing machines based on them. They lacked the power needed to support a slew of users, process vast quantities of data and all of the other tasks that big business needed. So, end of story. As production techniques improved, the major computer companies were more concerned with improving their existing product lines rather than searching out new markets. The mainframe and minicomputer businesses still constituted the vast bulk of the computing market and, with business now fully aware of the potential of computing, demands for greater storage and faster processing were the order of the day. In 1973 IBM still ruled the roost and it's premier offline storage product was the 3330-11 - a 400Mb mainframe disk drive. Costing a whopping $111,600, the drive offered unheard of storage but things would rapidly change. By 1980 $87,500 would get you 2.5Gb of storage (2.5Gb = 2560Mb). Memory was also starting to drop in price and in 1979 the company reduced it's price from $75,000 per 1Mb to $50,000 (as way of comparison, at the time of writing this (mid 2005), 1Mb costs about $0.11). As the 1970's progressed, the mainframes and minicomputers steadily got bigger, faster and started introducing multiple processors, but the underlying principles and technologies remained unchanged. It was the microcomputer industry that was undergoing the biggest changes and none of the big companies were interested. DEC's Ken Olson openly stated that he couldn't see any reason why anyone would want a computer in the home and this was not an entirely unsubstantiated statement given the disaster that had befallen Honeywell in 1969. Having released their 'Kitchen Computer', Honeywell were sure that every housewife would be keen to have a machine that could store recipes, tell you which dishes could be made from which ingredients and balance the chequebook. The fact that it took two weeks to program and cost a staggering $10,000 (in 1969) meant that the Neiman-Marcus chain shifted precisely zero of them, and the power of computing failed to infiltrate the home. This lack of interest from the likes of IBM towards the home computing was understandable and certainly suited the little startup companies, but also stopped home computing from really gaining a visible presence in the marketplace (outside of enthusiast circles). The Apple II went a long way to changing that and more than ruffled a few feathers on the way. IBM, like so many of the other 'big' manufacturers had paid little or no interest to the microcomputer but the success of the Apple II had propelled the little company into the big time and, mainly due to applications like VisiCalc, business users were starting to look towards the microcomputer to supply their computing needs. With it's low cost and small size, workers could, for the first time, have the same machine at home that they used in the office. By entering the business sector, Apple had just awakened the corporate beast...and the corporate beast had to get back into the game very quickly. Despite being forced into the game, IBM had already produced a microcomputer in 1975 - the 5100. With it's built in 5" screen and 16Kb of RAM it was a nice little machine but at $10,000 it was out of the league of most hobbyists and the machine quickly died away and left IBM distinctly unimpressed. In the face of the brand new microcomputer market though, IBM had to re-examine what it was selling and who it was selling it to. This new market wasn't big business or government funded facilities, it was ordinary people who didn't have thousands of dollars to spend on a 'frivolous' machine. The result was a new machine that would go on to dominate the market for years to come: the IBM Personal Computer (PC). After toying with the idea of outsourcing the project to Atari, IBM decided to stick with developing and producing the machine themselves and, on 12th August 1981, the first IBM PC hit the market. With an Intel 8088 processor running at 4.77MHz, 16Kb of memory and either one or two 160Kb floppy drives, the machine wasn't revolutionary but it did have the magic IBM logo on it and retailed for a not stupid $1565. Buyers who'd been wary of trusting little unknown firms were suddenly faced with the backing of the grand-daddy of the computer industry, and if you couldn't trust IBM, who could you trust? Just as important as the hardware was the software. Previously IBM had always produced their own operating systems and software but for the PC they contacted a little firm called Microsoft. Started in 1975, Microsoft adapted and sold a version of BASIC for the MITS Altair and their success persuaded it's founders (Paul Allen and a chap called William Gates (whatever happened to him ehh?)) that that there was money to be made in software. Microsoft offered to write a version of the BASIC programming language and IBM also offered them the chance to develop the operating system for the machine. This was new territory for Microsoft and they declined, instead pointing Big Blue in the direction of Gary Kildall, the developer of the (then) very popular CP/M operating system. Trying to setup a meeting with Kildall, IBM succeeded in contributing to one of the most amazing incidents in computing history. With Kildall having left for the day, IBM executives were forced to meet with his wife who then proceeded to refuse to sign a non-disclosure agreement. IBM left empty handed and returned to Microsoft offering them, again, the opportunity to develop the operating system. This time Microsoft agreed and quickly started working on a variation of QDOS (Quick and Dirty Operating System) written by Tim Paterson of Seattle Computer Products. QDOS had been a prototype operating system designed to work with Intel's new 8086 processor and would be ideally suited for the new IBM PC. Of course, in a cruel twist of fate, QDOS was based on Kildall's CP/M. Having paid $50,000 for the rights to QDOS (having kept the IBM deal secret), Microsoft produced their Microsoft Disk Operating System - MS-DOS. Gates then talked IBM into letting him retain the rights to MS-DOS, leaving the way open for other manufacturers to use it if they could strike a deal with Microsoft (as opposed to IBM). "As it'll only run on IBM hardware then where's the risk?" thought Big Blue. Goodbye To The Command Line With the IBM PC having made a big splash (despite ridicule from market leaders Apple who had even tried to bait the newcomer with full page adverts 'welcoming' them) the microcomputer was becoming something of a must have item. This wasn't just in the home market but also in the business sector. Users at all levels felt happy that they had their own machine and didn't have to share it with anyone else (as most users still did when it came to mainframes and minicomputers - not that it actually made any difference due to the sheer power of these behemoths). They'd happily switch on their computers, type in the appropriate commands and away they'd go. Elsewhere some people thought differently though. Way back in the late 1960's a group led by Doug Englebart had demonstrated a project called NLS (oNLine System) that used graphics, video, audio and a strange box with a wire coming out of it. NLS had quickly died following it's initial presentation but its influence had carried on into Rank Xerox's Palo Alto Research Centre (PARC). PARC was a hot bed of the brightest computer scientists who had the simple task of developing new technology. They didn't have to make it practical, affordable or realistic - they just had to see what was possible. The biggest breakthrough (conceptually) had been the development of SmallTalk, a development system that allowed a new approach to how computers were used, to be implemented. Instead of treating the machine as a command line, the PARC researchers expanded on the approach developed by the NLS team - the computer wasn't a tool but was a virtual office. Files were stored in a filing cabinet, letters arrived in an 'inbox' on your desk (and were sent out through the 'outbox') etc. This way of thinking did not lend itself well to the traditional command line (where users had to type in a command, let it run and then enter the next one) - there was no way to do things concurrently, switch between tasks quickly or share data/information easily. Englebart had proven that it could be done and PARC refined it, expanded it and produced the first graphical user interface that was actually put into practice (Englebart's NLS was good and had been demoed...but never actually got any real practical use). Started in 1972, the machine that would become the Xerox Alto was designed as a personal research machine that would quite happily allow data to be shared across a network and documents to be printed, but still meet all of the needs and requirements of the individual user, providing offline storage and plenty of processing power. With it's portrait display and desktop keyboard and mouse, the Alto was great but what wasn't so good was the enormous processing unit that went along with it (about the size of a large bedside cabinet) and the estimated $32,000 price tag (and that's 1979 money). This wasn't a concern at PARC though so the machine was heralded as a success and quickly spread around the whole complex. It also spread elsewhere though and word of mouth soon had details of the Alto and PARC's work spreading throughout the IT industry. Few were allowed in to see what all the fuss was about but Apple's Steve Jobs was one of the few (despite the fact that it took some persuasion by key Apple employees to get him to go). What Jobs saw that day would change mass computing forever. Apple had already started work on a new machine, Lisa, (a more business orientated machine to compliment the Apple II) but its design stipulated a command line. This was standard at the time so didn't cause any problems. Following his visit to PARC though Jobs insisted that Apple develop a graphical user interface (GUI). The GUI allowed users to forget about cryptic commands and instead use the computer as the virtual office that the NLS team had been striving for. Eventually Apple's Lisa saw the light of day in 1983 and it was a revelation...to most. Although the PARC team didn't have to concern themselves with making practical products, Xerox did and, using the technology developed at PARC, released the Xerox 8010 Star in April 1981. Sporting a GUI, the Star was a technological marvel but gained little interest, mostly due to its high price tag. The Lisa, although still expensive at $10,000, had the Apple marketing machine behind it and was the buzzword on everybody's lips. Lisa was revolutionary to the market but it didn't sell well (mainly due to the price tag) but Apple's Macintosh product (which shared the GUI - if not the same one) addressed many of the problems and retailed at a quarter of the price. Sales were significantly better and the potential market for microcomputers exploded (even if it would take several years for the GUI Macintosh to reach the same level of sales as the 'archaic' Apple II). Even given the swish plastic case of the Apple II, users still had to have a modicum of intelligence and still had to remember cryptic commands - the GUI solved all of that. Now users could simply flick a switch and start moving the mouse about (that funny box with a wire coming out of it), clicking on little pictures and never having to worry about whether to use BLOAD or BRUN - it just worked. Even the big mainframes weren't doing this kind of thing. The Greatest Form of Flattery IBM's PC had given Apple some competition but the GUI had allowed the fruit flavoured company to take the lead once again (even if its products were that bit more expensive). GUIs were, undoubtedly, the way to go and Microsoft began work on its own GUI for the PC, announcing the fact in 1983 - Microsoft Windows v1.0 wouldn't ship until 1985 though. Despite the lengthy development process, Windows was not an ideal replacement for Apple's Mac OS and it wasn't until Windows v3.0 in 1990 that the product really became the must have program for PCs. Apple hadn't stood still and the pair would be forever updating and re-inventing for many years to come (as well as being embroiled in bitter arguments about who'd stolen which ideas from who). It wasn't just software that was open to 'flattery' as IBM quickly discovered. Founded in 1982, a little startup firm named Compaq had just one goal: create a microcomputer that could run IBM PC software. The goal was easy but getting there wouldn't be. Although IBM had used 'stock' parts and allowed Microsoft to retain control of MS-DOS (therefore allowing any company to license it) the PC had incorporated a special BIOS (Basic Input/Output System) chip that controlled how the machine worked. Without this magic chip, MS-DOS was pretty useless as no-one could create a machine capable of doing what the PC did (which probably explains why IBM were happy for Microsoft to retain ownership). Compaq realised that if they directly copied the BIOS chip then they'd be in serious trouble. The solution was to create two 'clean' rooms. In one room the PCs BIOS chip would be dissected and reverse engineered, while in the second, a 'new' version would be built that could do the same things. The distinction was vital as it allowed Compaq to prove that it hadn't actually copied the chip itself, merely created a chip that could do the same thing. A small distinction but enough to allow Compaq to release the first IBM PC compatible, in just the same way that Gene Amdahl had released IBM 360/370 compatible machines in 1975. With the door now opened, 'clone' manufacturers started to flood the market and IBM quickly found themselves in a dangerous position where it was being rapidly priced out of the market. Despite the fact that these new manufacturers weren't IBM, the fact that they had 'IBM Compatible' on the case was more enough for customers on a budget. The big winner in all of this was Microsoft who, despite IBM's 1981 'trump card' (in the shape of the BIOS chip) now found themselves in the dominant position. Any manufacturer could now legally duplicate/produce the all important BIOS information but they were all reliant on MS-DOS (and later Windows)...and that could only come from one place. Into The New Era...And Beyond Where hardware had once been the important factor, by the 1990's software had become the real money making industry. While Apple closely guarded it's hardware and software (and still does, despite a brief spell where other manufacturers could license the all important software) the PC became nearly totally controlled by Microsoft (various Linux implementations are starting to gain widespread use, sometimes even with government support) despite the fact that the Redmond firm didn't actually produce hardware. Mainframes and minicomputers quickly returned back to niche installations in the wake of the microcomputer and, while IBM still does a nice line in them (which is has to, having recently withdrawn totally from the desktop market) the microcomputer now rules the roost in terms of user hardware. Technology has moved on enormously and today's machines (even desktop PCs) can perform hundreds, thousands and even millions of times more operations per second than the likes of UNIVAC and ENIAC (although a project was undertaken in 1997 to create ENIAC-on-a-chip). Machines that once filled entire offices now come in boxes smaller than a shoebox, produce next to no heat, consume no more power than a light bulb, and certainly don't take three weeks to program using a plugboard and a team of mathematicians. Memory and disk storage prices have plummeted and machines no longer cost millions of dollars, a PC now coming in for just a few hundred. With the advances in new markets slowing (computers have penetrated nearly every level of society so there's no-one new left to sell them to) and system architectures having fallen into a 'faster rather innovative' vein, the 1990's and the 21st century have seen the explosion of networking and the sharing of information. Konrad Zuse might have had no idea of where his 1930's relay based machines might lead to, and it's unlikely that anyone could have predicted the impact and uses that computers would eventually have, but one thing is certain: Howard Aiken was about as far off the money as you could get when he predicted that the world would only need 6 computers. Want To Know More? This article was intended to be a quick look at the history of computing and naturally there's far more to the subject than just the insane ramblings of a Mac fanatic. The internet is jammed packed full of useful (and sometimes, not so useful) information but in the process of researching this subject I found that, as per usual, nothing could beat the good old reference books. So if you're looking to find out a bit more, you could do a lot worse than trying to track down some of the following: Electronic Brains - Mike Hally (Granta Books, ISBN: 1-86207-663-4) A Computer Called LEO - Georgina Ferry (Harper Perennial, ISBN: 1-84115-186-6) Early British Computers - Simon Lavington (Manchester University Press, ISBN: 0-7190-0810-7) Computers An Illustrated History - Christian Wurster (Taschen, ISBN: 3-8228-1293-5) Fire In The Valley - Paul Freiberger & Michael Swaine (McGraw Hill, ISBN: 0-07-135892-7) Revolution In The Valley - Andy Hertzfeld (O-Reilly, ISBN: 0-596-00719-1) Digital Retro - Gordon Laing (Ilex Press, ISBN: 1-904705-39-1) Insanely Great - Steven Levy (Penguin, ISBN: 0-14-029117-6)

Site designed and maintained by TheNeil. While all content is checked and updated regularly, the author cannot be held responsible for any broken links, incorrect information or damage caused to hardware or software. Comments, contributions and criticism always gratefully received. Free Website Stats See that? That's the number of fools that have found their way here Site Last Updated: 11/04/2009 11:26:21