As the title says, I have no idea what these things are or how to get or use one. Can I receive a simple explanation or links to one? I'm a computer engineer so I have very little physics/mechanical engineering background.
A strain gauge is a device used to measured the strain (change in length as a proportion of the original length) in an object as a result of an applied load. Most strain gauges are designed to measure strains in only one direction.
How it works:
A common type of strain gauge consists of thin metallic foil cut into a pattern such that most of the current flow is oriented lengthwise along the object you are attaching the gauge to. As the object is subjected to a tensile strain, the gauge elongates longitudinally. This causes the lengths of conductor parallel to the strain to become longer and (as a consequence of Poisson's ratio) thinner.
For a fixed resistivity, the resistance of a conductor is inversely proportional to the cross sectional area (which is decreasing) and directly proportional to the length (which is increasing). This can be expressed as the relation:
Resistivity ($\rho$) is a material property, so it remains constant. Thus, we know that the resistance of the gauge is increasing.
Measuring this change in resistance allows us to find out the applied strain, by the relation:
where GF is the gauge factor (constant for a gauge). The gauge factor can be determined by measuring the change in resistance caused by a known strain.
I worked with strain gauges in a lab. They are snakelike traces etched onto a plastic substrate about the size of a dime, and you glue them onto your test surface with special glue. You solder leads on and wire it back to a wheatstone bridge circuit. I think the standard is 350 ohms, so you would have three fixed resistors and your strain gauge. If you want to get fancy you can get an extra strain guage and glue it down to a dummy piece of test material for temperature compensation, but plain resistors are fine for the other branch of the wheatstone bridge. It works out that with a 2-volt excitiation, I think, microstrains = microvolts, so you want a really good voltmeter that displays five or even six decimal places for accurate work.
(A microstrain is a one-part-per-million deformation.)
To add to Marty's answer up there, a strain gage (guage sp?) is typically used to measure applied load. You might wish to check out the wikipedia article on load-cell
The strain gauge has known electrical resistance that changes when it is strained (load applied to it). This change is a function of the load; hence the applied load can be determined by measuring the electrical resistance on loading. The principle is used in weigh-bridges at toll-booths, railway sidings, docks etc to translate the mass of a vehicle/body into a form that may be directly read by a computer.
(We used it to measure coal brought by railway wagons to the dump-chute by applying the output from the bridge to an ADC which was then processed by the application )
This seems more like a practical mechanical engineering question than a physics question. There are a great variety of strain-sensors available commercially. What you need really depends upon your application. There are entire books written on the subject. An appropriate sensor for measuring the strain of a bridge or a concrete dam would be quite different than a sensor intended to be used inside a transmission electron microscope. If you need a practical answer about strain-gauges, your best source of information may be strain-gauge manufacturers.
There are some very interesting (IMO) physics questions that arise if you need an unconventional strain-gauge for a specialized application. Strain is essentially displacement divided by a length. I have seen original strain-gauge designs developed (with input from physicists) based on resistance, induction, lasers, fiber optics, and piezoelectricity to measure strain in everything from living muscle to glaciers.
The is also much opportunity for work here for computer scientists, because all the information returned by strain sensors must be numerically processed to be useful.