Ideals and Unique Factorization
Here is the shock that nearly sank the early attempts on Fermat's Last Theorem:
unique
factorisation — the bedrock of ordinary arithmetic — can fail in a
ring of algebraic integers.
The fix, due to Kummer and Dedekind, was to invent a new kind of object: the ideal.
The failure
Work in \mathbb{Z}[\sqrt{-5}]. The number
6 factors in two genuinely different ways into irreducibles:
6 = 2 \cdot 3 = (1 + \sqrt{-5})(1 - \sqrt{-5}).
All four factors are "prime" in the sense of being unbreakable, yet the two factorisations share
none of them. Unique factorisation is simply false here. Without a repair, every argument that
secretly relied on it collapses.
The rescue: factor ideals, not numbers
Kummer's insight: factorisation is restored if you factor not the numbers but the
ideals they generate. An ideal is a set of ring elements closed under addition and
under multiplication by anything in the ring — think of it as "the multiples of a number", but
promoted to a first-class object that can exist even where no single generating number does.
In the ring of integers of any number field, every nonzero ideal factors uniquely
into a product of prime ideals.
How it heals the example
The two clashing factorisations of 6 come from
refining each number into prime ideals — and at that finer level the two refinements agree:
(6) = \mathfrak{p}_2^{\,2}\, \mathfrak{p}_3\, \mathfrak{p}_3', \quad\text{with}\quad (2) = \mathfrak{p}_2^{\,2},\ \ (3) = \mathfrak{p}_3\,\mathfrak{p}_3'.
The numbers 2, 3, 1\pm\sqrt{-5} were only coarse groupings of
these true prime-ideal atoms. Unique factorisation was never really lost — we were just factoring
the wrong objects. This single idea, born from Fermat's Last Theorem, founded modern algebraic
number theory.