086

The Computer History Museum, Mountain View, CA

37° 24′ 51.74″ N, 122° 4′ 36.54″ W

Silicon Valley’s Finest

Almost all major science museums around the world have sections tracing the history of the computer, but they all pale in comparison to the Computer History Museum in Mountain View, California. The museum is housed in the old Silicon Graphics headquarters (itself an iconic Silicon Valley building), which the museum bought in 2002. Silicon Graphics was once a maker of high-performance computers for computer graphics that made possible 1990s films like Jurassic Park; its former headquarters is located on North Shoreline Boulevard next to Charleston Park, and the design and architecture of the building speak of the extreme wealth of Silicon Valley prior to the end of the Internet boom.

The museum is a work in progress, with as much information available online through its excellent website (http://computerhistory.org/ ) as in the building itself. But unlike many computer science museums (including the worth-avoiding Tech Museum of Innovation in San Jose), the Computer History Museum brings computing to life through exhibits that are explained both in general terms for the casual visitor, and in glorious detail for the gearheads. You’ll get the most out of your visit by going on a docent-led tour. These tours are not canned recitals from a script, but truly bring to life the machines in the collection. In addition, everything is free—the museum is supported by donations of money, machines, and software. Consider making a donation before you leave.

A tour of the 500 pieces of the collection on display is a good way to get oriented. The collection includes bits of 1954’s SAGE (Semi-Automatic Ground Environment; see Figure 86-8), which was built as part of the U.S. air defense mechanism during the Cold War. SAGE weighed 300 tons, and incorporated just one comfort item for the operator looking out for Soviet bombers: a cigarette lighter. Some parts of the SAGE system were still in use in 1983, with replacement parts purchased from Soviet-bloc countries because they were unavailable in the U.S.

Figure 86-8. SAGE on display at the museum; courtesy of Flickr user Jurvetson

The 1960s collection includes the industry-dominating IBM 360 mainframe computer with its spinning tape drives. 1960s attitudes are on display with the Honeywell Kitchen Computer sold by upscale department store Neiman Marcus in 1969: the advertising tagline reads, “If she can only cook as well as Honeywell can compute.” The computer incorporated a cutting board and was preprogrammed with a collection of recipes; it’s not clear if any were sold.

The museum has an excellent collection of supercomputers including a CRAY-1, CRAY-2, and CRAY Y-MP, as well as a Connection Machine CM-1. More recent computer history covering the personal computer revolution includes exhibits of the Altair 8800 (the machine that made Bill Gates leave Harvard and start “Micro-Soft”), an Apple I, the original IBM PC, the almost-but-not-quite portable Osborne 1 Portable Computer (only 24 pounds!), the Commodore 64, the Apple Lisa and Macintosh, and a French Minitel. Along with the machines, the museum has also preserved software including Microsoft Windows 1.0, and peripherals such as the 1978 Dover Laser Printer.

Pre-computer history is covered with a collection of mechanical devices (such as slide rules), a Hollerith Census Machine from 1889 that used punch cards, and a Jacquard Loom (also using punch cards for weaving).

A demonstration well worth seeing is the operation of the only working DEC PDP-1 in the world (this happens twice monthly, on a Saturday afternoon). The PDP-1 was first built in 1959 (the example on display was made in 1963 and restored to full working condition in 2004) and was the first commercial, interactive minicomputer. Data is fed in using a punched or magnetic tape, and the user interacts with the machine using a typewriter/printer combination, or the screen and lightpen. The PDP-1 has a memory of about 9 kilobits, but that was enough for serious scientific applications, playing music, and a game: Spacewar!.

Spacewar! is demonstrated along with the PDP-1, with two players flying ships orbiting a star, and trying to shoot each other while avoiding the star’s gravity.

Practical Information

The Computer History Museum is located on North Shoreline Boulevard in Mountain View, California, and is accessible directly from US-101 at the Shoreline Boulevard exit. It can be reached from the Mountain View Caltrain Station via the Caltrain Shoreline Shuttle. Full details can be found on the museum’s website, http://computerhistory.org/.

The Transistor and NAND Logic

Fundamental to the electronic computer revolution is the transistor. A transistor is a small electronic device that can act as an amplifier (they were used to make the “transistor radios” of the 1950s and 1960s, which were the first truly portable music players) or as a switch controlling the flow of electricity in a circuit.

A semiconductor is a material (or more commonly, a combination of materials) that is neither a conductor of electricity nor a complete insulator. The first semiconducting material (silver sulphide) was discovered in 1833 by the great British scientist Michael Faraday (see Chapter 75). Faraday discovered that when heated, the resistance of silver sulphide was lowered and the compound conducted better (and the effect reversed as it cooled).

In a transistor, the amount of electricity flowing through a semiconductor, typically silicon, is controlled electrically. This control means that the transistor can operate as either a switch (turning on or off the flow completely), or an amplifier (since the amount of flow is proportional to the voltage applied).

A simple silicon transistor, called an NPN, is used to make the electronic switches that are the building block of computers. The NPN transistor consists of three pieces of silicon known as the base (often written B), collector (C), and emitter (E), as shown in Figure 86-1.

Figure 86-1. Parts of an NPN transistor

The base of the transistor is connected to the P; this area is silicon that has been modified in a process called doping, where the silicon is deliberately exposed to a chemical such as boron, removing free electrons. The P stands for positive, reflecting the dearth of electrons. It is the base B that controls whether the transistor switch is on or off.

The collector and the emitter are made of N-doped silicon; this is silicon that has been exposed to a chemical like arsenic to increase the number of mobile electrons available to pass a current. The N indicates that this region is filled with negatively charged particles: electrons. With no voltage applied between the base and the emitter, the base material acts as an insulator and the transistor switch is off. Once a sufficient voltage is applied between B and E, electrons are able to move between the collector and the emitter and the switch is on.

NPN transistors are usually drawn using the symbol in Figure 86-2 showing the emitter, base, and collector. The base, on the left, controls the flow of electricity between the collector and emitter.

Figure 86-2. A transistor

Computers use binary logic consisting of just 0s and 1s for all calculations (see section on Binary; page 218). This is partly because binary is easy to work with electrically: no current flowing is 0, current flowing is 1. A transistor is simply a switch that’s on when current is flowing through the base (when it’s handling a binary 1) and off when there’s no current (a binary 0).

Using transistors, it’s possible to build up the basic logical functions needed by a computer. Computers combine two or more binary numbers using so-called logic gates to make decisions. For example, an AND gate expresses the concept “If inputs A and B are both 1 then output a 1, otherwise 0”; an OR gate expresses “If inputs A or B or both are 1 then output a 1, otherwise 0”; and a NOT gate expresses “If input A is 0 then output a 1 and vice versa.” From these simple, logical decisions, it’s possible to build entire chips.

But one such logic gate, called a NAND, trumps all the others because it’s the only one necessary. NAND means NOT AND and expresses the concept “If either of input X or Y is 0 then output a 1, otherwise 0.” It’s the opposite of AND: where AND would say 1, NAND says 0. It turns out that any other logic gates (AND, OR, NOT, and some other more exotic ones) can be made just from NANDs, and a NAND gate can be made from two transistors (Figure 86-3).

Figure 86-3. A NAND gate

If both of the inputs A and B are turned on (a binary 1), then both transistors turn on, and no current flows from the output because it’s flowing through both transistors: the output is 0. If either of the inputs is turned off (a binary 0), then one of the transistors will be off, and the current will flow through the output, giving a binary 1.

NAND gates are usually drawn with the symbol inFigure 86-4.

Figure 86-4. Symbol for a NAND gate

Making the other logic gates from NAND gates is fairly simple: to make a NOT gate (which outputs 1 when its input is 0 and vice versa), just join the inputs A and B together (Figure 86-5).

Figure 86-5. A NOT gate from a NAND gate

From that, it’s easy to make an AND gate by taking a NAND gate and flipping its output using a NOT (since NAND is NOT AND, and NOT NOT AND is just AND). See Figure 86-6.

Figure 86-6. An AND gate from two NAND gates

OR is the most complex, requiring three NAND gates (the first two work as NOT gates, flipping the inputs A and B). See Figure 86-7.

Figure 86-7. An OR gate from three NAND gates