[rakmanenuff]

it's theoretically possible to design such a software, but it would
probably get it "wrong" on maybe... 1 out of 10 tracks.

if EQ was the only problem it'd be easy, but there is the whole thing
of how "loud" a track sounds to the person listening, and that can be
totally different from what the meters are showing. there are so many
possible variations, imo the software just wouldn't be able to
anticipate it all. but maybe that could be corrected over a longer
time period of trial and error.

if you're dj'ing i guess you just have to know your record collection
and use the EQ. or you could use AUDACITY to do your own masters of
the most problematic tracks beforehand.