proper exposure what is that?
Took you long enough, but at least you've finally acknowledged your ignorance. Congratulations! It's the first step towards knowledge.
In the digital realm, a proper exposure is one in which an image of an 18% gray card results in values in the green channel of the raw file of 18% of the maximum value that can be recorded. With a 16-bit file, values range from 0 through 65,535; 18% of that is 11,796.
The red and blue values will be less than the green values in all but the weirdest light. Typically, red will be a bit more than a stop relatively underexposed and the blue will be about 2/3 stop underexposed. White balancing consists of multiplying the figures from the two underexposed channels so they equal the figures from the properly exposed channel.
Digital exposure compensation is the same basic idea, but applied equally to all three channels. One stop of digital push is accomplished by doubling all the figures in the raw file. One stop of digital pull is done by halving all the figures.
After white balancing and exposure adjustments, a gamma curve is generally applied (with 2.2 being typical of most working spaces). Next, if it's a colorimetric workflow, the ICC profile for that camera is applied. Much more common, however, is that this is where the "special sauce" gets applied. Canon calls them, "picture styles." At the very least, an S-curve is applied (on top of the gamma curve) to boost contrast in the midtones (at the expense of highlight and shadow contrast, of course) and different parts of the spectrum are selectively saturated by different amounts. (It's always added saturation because the camera's native color response is so much larger than any working space). After that, whatever other types of additional modifications (more or less contrast, shadow boost, whatever) are up to you.
In scenes with reflective highlights or the light sources themselves, you generally want to intentionally underexpose the image because otherwise those highlights will oversaturate the sensor and one or more of the channels will record the maximum value for those parts of the image. Underexposing allows those values to fall in a range that can be recorded by the sensor; digital push is then applied to most of the image to restore it, but a compressed push is applied to the highlights -- if you didn't compress the boost in the highlights, they'd wind up just as blown as if you had properly exposed the image in the first place.
Scenes like that are so common in typical shooting that manufacturers intentionally design their meters to underexpose by about a stop. The onboard raw development engine that produces JPEGs applies a corresponding amount of compressed digital exposure compensation, and the various popular raw development engines are smart enough to do the same. There are, however, many tools out there that will let you examine the raw data directly should you choose to play around.
(That scene of the shed in my back yard where I moved the camera's exposure bug six stops to the left, that resulted in an almost entirely black exposure with just hints of detail? The one which came out almost entirely free of noise except for a small amount in the deepest shadows after normalizing the exposure? Considering the meter's built-in underexposure, it was really about seven stops underexposed.)
And now you should understand what proper exposure is.