These are easy. Most buttons/switches have either two or three terminals. When you press the button or flip the switch contacts between the terminals are made and/or broken. The details of what happens depends on the exact configuration of the button/switch. Some of the variables are: number of switch positions (poles), number of switches (throws), normal state (open or closed) of each positions.
A single pole, single throw (SPST) switch is the simplest kind. A double pole, single throw switch is slightly more complicated. They look like this:
To use a SPST switch, you attach a voltage source to one terminal of the switch. When you flip the switch a connection is made between the two terminals, which means that current can now flow from one terminal to the other. That doesn't sound very exciting, until you remember that a microcontroller has digital input pins. By hooking up a button between a +5v voltage source and the input pin on a microcontroller you can sense whether the button is pressed or not. When it's not pressed, the microcontroller will see no voltage on its input pin (a "0"). But when it's pressed it'll see +5v (a "1"). (Sometimes you need a "pulldown resistor" on the pin so that the pin really sees ground when the button isn't pressed. It depends on the microcontroller.)
There are lots prefabricated buttons and switches available. You can also make your own pretty easily -- all you need is two pieces of conducting metal! Pretty much any instrument or device that has a make/break contact mechanism (clarinet, piano, hi-hat, phone, door, foot on the floor, window, etc.) Can be wired up as a button/switch sensor.
potentiometer
A potentiometer (or pot) is a simple knob that works as a voltage divider with two variable resistors. Small potentiometers are called trim pots. They come in many shapes and size and value ranges. A potentiometer will be labeled with its total resistance. Potentiometers are either linear or logarithmic ("audio"). When using a linear pot Vout will change linearly as the knob is turned. When using a log pot Vout will change logarithmically as the knob is turned.
A pot has three terminals (or legs), which correspond to Vin, Vout and ground in the voltage divider diagram in the previous section. You hook your +v up to the first leg (Vin), ground up to the third leg (ground) and use the middle leg as your output (Vout). As you turn the knob on the pot the voltage on the middle leg will get bigger/smaller.
Potentiometers are very useful for setting parameters, like the speed of a process or the brightness of a light. They're also good for measuring the rotational angle of an object about an axis. Log pots are good for controlling non-linear parameters like pitch; they give you more resolution (change slowly) in the bass and less (change quickly) in the treble.
pressure sensor
A pressure sensor is generally a variable resistor that changes resistance based on how hard it is pressed. They are also known as Force Sensing Resistors or FSRs. They come in many shapes and sizes, from tiny little dots to long strips. They're pretty flexible, and can usually be cut to size.
Often you use an FSR as the variable resistor in a voltage divider circuit. By using one FSR and one fixed resistor and hooking Vout to an A/D converter you can give the microcontroller some idea of how much pressure is being put on the FSR. If you don't have an A/D converter handy, there's a simple RCtime circuit that you can use to do more-or-less the same thing.
FSRs can be used in many places that a button is used, with the advantage that you can not only tell that the button is being pressed, but also how hard. They're used a lot in things like velocity-sensitive keyboards, but are also good for use in floor mats, on the sides of instruments, on finger tips...Basically anywhere there's going to be a force that you want to measure.
light sensor
The most common light sensor is the CDC photo sensor. It works like a variable resistor with a value that changes based on how much light it receives. Other types are photo-transistors and photo-diodes.
Photo sensors are used in much the same way as pressure sensors. If you hook one up to a microcontroller via a voltage divider (or RCtime) circuit, the microcontroller can sense how much light is falling on the device.
beam breaker
A beam breaker is a non-physical switch or trigger. The simplest way to make one is to aim a cheap laser pointer at a photo-sensor. When the beam is broken, the resistance of the photo-resistor changes. If the photo resistor is hooked up to a microcontroller, it can sense whether the beam is intact or broken. So simple!
proximity sensor
There are a variety of ways to implement proximity sensing. The most common are infrared and ultrasound. They're both usually used as packaged devices that hook up to a microcontroller via one or more digital i/o lines.
An infrared sensor works by sending out pulses of infrared (invisible) light. It then tries to detect reflections of that light from nearby objects. If it detects a reflection then it assumes that there's an object nearby and puts out a digital "1". Otherwise it puts out a "0". There are some tricks to get a rough distance measurement out of an infrared detector, but most often they're used for simple object detection.
An ultrasonic proximity detector works by putting out ultrasonic (inaudible) pulses of sound. It then measures the length of time it takes those pulses to hit nearby objects and return as echos. The longer the time, the farther away the object. Ultrasonic proximity detectors are good for measuring short distances (maybe a couple yards). It's usually up to the microcontroller to do the counting and distance computations; the detector takes care of sending and detecting the ultrasonic pulses.
accelerometer
Accelerometers are cool. They're small ICs with tiny little suspended weights embedded in them. By measuring the force of gravity/inertia on those weights the device can emit a digital signal that tells you about its orientation in space.
Like proximity sensors, accelerometers are hooked up to a microcontroller via a few digital i/o lines.
microphone
Most microcontrollers aren't fast enough to be useful for recording sound via a mic. But they can do some basic things like listen for suddens sounds or listen for silence. They can also be used to control sound recording devices.
contact mic
These are made from piezo elements. A piezo element is made from a curious type of ceramic material that gives off electricity when it's deformed (this is called the piezo-electric effect). The voltage created is (roughly) proportional to how much the element is deformed. This property lets it be used as a contact mic. You can scream all you want into a contact mic and it won't hear you. But rub it against something, and it will wail! There are other types of piezo materials, like piezo rubber, that could be fun to play with.
