[BrianG]

Improper fractions are kinda necessary when doing proportions as Metallover indicated. How else would you do the cross-multiplication? I suppose you could convert the "proper" fractions to a floating point decimal, but you can easily lose accuracy due to rounding (especially if that answer is used in another calculation).

Of course, it depends on what the person gets into later in life. Probably the most common thing I do in programming is derive formulas from data. It's not especially difficult, but to do stuff like that, one needs to understand all the basics.