When the eye looks at a scene with dark shadows, and bright highlights, it will adjust to the lighting conditions of the point of focus (IE pupils dilate or contract to let more or less light in, depending on which part of the scene you are looking at). The camera sensor is not able to do this, because the 'dynamic range' - the amount of light-to-dark it can capture - is much less than that of the subject. So, do you 'expose for the highlights', so that the sky looks right and the rest is a silhouette? Or do you 'expose for the shadows', and make the sky a bleached out mass of whiteness, with no detail there whatsoever?
What you do is both, and take a middle one too. For example, 3 exposures are taken of the same subject, from the same point (on a tripod!) but the exposures are deliberately underexposed, normal, and overexposed. Between the 3 exposures, you have the detail you need, but in order to get the correctly exposed bits of each scene into a single well-balanced image, they need to be merged. This process is called HDR and involves some software jiggery-pokery (photographic technical term).
The thumbnail images on this page are all HDR examples. (When I have time, hovering and clicking will bring up larger size examples)
© Ian Shewan Photography