This animation is meant to clarify what "coherence length" means.
The top wave represents the kind of light produced by most sources--incoherent. Some sources (e.g., the sun) produce white light--all wavelengths. We assume here that we've used filtering to concentrate on only one wavelength. Some common light sources, such as mercury vapor lamps, are better at producing only certain wavelengths, but they are still incoherent. So what does it mean for light of a given wavelength to be incoherent? Back to the top wave: It has a regular wavelength, but after some number of regular waves, another "piece" starts, whose wavelength is the same, but which picks up at a different, random point in the cycle. Natural light is like this, for it is produced by a succession of events--charged particles moving about--which know nothing about each another, and so, while perhaps producing the same wavelength, don't do so in phase with others.
The middle wave is the top wave progressively shifted to the left. (If you move the slider quickly you'll get the illusion of its moving to the left.) In the first frame it is not shifted at all, and each successive frame shifts it one half wavelength leftward.
The bottom wave is the algebraic sum of the top and middle waves. Since the middle wave is not yet shifted, the first frame shows an algebraic sum that is simply twice the top wave. After that, we begin to see the important point. When the middle wave is shifted one half wavelength, destructive interference is seen almost everywhere. Only at the "joints" between pieces is it messed up. Another half-wavelength shift puts the waves almost back in a constructive situation. Another half-wavelength gives near destructive; another near constructive. However, because these waves are not regular for very many wavelengths (a coherence length of 5±1 wavelengths was chosen), after not too many half-wavelength shifts, it becomes hard to tell what is supposed to be constructive from what is supposed to be destructive. From then on, the shift is essentially irrelevant; the average height of the algebraic sum is unchanged; the interference has been lost.