Lately I was thinking about what the normalizing feature in most of the ripping proggies (EAC, CDEx) really do. I always used this feature to prevent the tracks from having clipping sections but doesn't normalizing also take away the dynamics in an audio track ?

Yup. That's why many of us prefer to use mp3gain to do lossless normalising after encoding (without scaling).

But doing this after the encoding process can't ever be as accurate as doing this before the encoding, that's what my logic tells me, right ?

You should try to understand how mp3gain works first

http://www.geocities.com/mp3gain