Magma's commutative algebra and algebraic geometry machinery spans roughly fifteen chapters of the Handbook. Rather than diving deep into one of them, this lesson is a map: what each area is for, how the areas connect, and where you would go next if you needed to actually compute something. Think of it as a table of contents with commentary, not a manual.
Groebner bases and ideals: the computational core
Everything else in this survey ultimately rests on Groebner basis theory. Given an ideal in a polynomial ring, a Groebner basis (with respect to a monomial order) is a generating set with special reduction properties that make ideal membership, elimination, and dimension computations effectively decidable. Magma's GroebnerBasis intrinsic computes one directly:
P<x,y,z> := PolynomialRing(Rationals(), 3);
I := ideal<P | x^2 + y^2 - z^2, x - y>;
GB := GroebnerBasis(I);
The ideals chapter builds directly on this: once you can compute a Groebner basis, you get primary decomposition, radicals, saturation, quotients, and tests like IsPrime or IsMaximal for ideals in multivariate polynomial rings. Together, Groebner bases and ideals are the engine room — nearly every other chapter in this survey calls down into them at some point.
A close cousin is localized polynomial rings: instead of working in the full polynomial ring, you localize at a maximal ideal (typically the origin) to study local behavior — germs of functions, local rings at a point of a variety, multiplicities, and tangent cone computations. This is the natural setting when a question is really about what happens near a point rather than globally.
Affine algebras and modules
An affine algebra is a quotient of a polynomial ring by an ideal, i.e., the coordinate ring of an affine variety. This chapter gives you the ring-theoretic view of a variety: computing its dimension, singular locus, normalization, and integral closure, and testing properties like being Cohen-Macaulay or Gorenstein — all through ideal-theoretic means. It's the bridge between "just an ideal" and "a geometric object with structure."
Modules over polynomial rings generalize ideals (an ideal is just a submodule of the ring itself) to genuine modules: free modules, submodules, quotient modules, syzygies, and free resolutions. This machinery underlies invariant theory, sheaf cohomology, and any computation that needs more than a single ideal — for instance computing a minimal free resolution to read off Betti numbers.
Invariant theory and differential rings
Invariant theory studies the ring of invariants of a group acting linearly on a polynomial ring — the subring of polynomials fixed by every group element. Magma computes generators of the invariant ring (via Molien series and constructive methods) for finite and some infinite reflection groups, which is the classical starting point for classifying orbits and quotient varieties.
Differential rings move commutative algebra into a different direction entirely: rings equipped with a derivation, supporting differential polynomial rings and differential ideals. This is the algebraic backbone for working with differential equations symbolically — differential Galois theory, D-modules, and related computations live here.
Schemes and sheaves: the geometric reboot
The schemes chapter is where the survey turns from "rings with properties" to "spaces." A scheme in Magma is built from an ambient affine or projective space plus defining equations, and you can ask for dimension, irreducible components, points over a field, intersections, and morphisms between schemes. It is the general-purpose object underneath curves, surfaces, and toric varieties alike — most of them are implemented as specializations of the scheme machinery.
Sheaves sit on top of schemes: coherent sheaves of modules (twisted structure sheaves, ideal sheaves) let you compute cohomology groups, which in turn give invariants like arithmetic genus and Riemann-Roch data. If schemes are the spaces, sheaves are how you extract numerical invariants from them.
Curves and surfaces
Curves — one-dimensional schemes — get an entire chapter of dedicated support: plane and space curves, singularities, resolution of singularities, genus, and rational points, tying together classical algebraic geometry with number-theoretic questions. Two companion chapters go deeper on specific fronts. Resolution graphs describe the exceptional configurations produced when resolving curve or surface singularities by blowing up, encoding how the resolution process unfolds combinatorially. Algebraic surfaces extends the two-dimensional case with surface-specific invariants and constructions (Del Pezzo surfaces, quotient singularities, minimal models). And Riemann surfaces takes the analytic viewpoint — computing period matrices, homology, and analytic invariants of curves over the complex numbers, complementing the purely algebraic curve theory.
Graded rings and toric varieties
A graded ring is a ring decomposed into graded pieces (think: a polynomial ring graded by total degree, or more exotic weighted gradings). This is the natural home for projective and weighted-projective varieties, since Proj of a graded ring is exactly how projective schemes get built from equations.
Toric varieties are a special, highly combinatorial class of varieties built entirely from lattice and fan data rather than explicit equations — a toric variety is described by a fan of cones in a lattice, and geometric properties (smoothness, completeness, the class group) can be read off combinatorially. They are a favorite testing ground precisely because so much geometry reduces to lattice combinatorics.
Putting it together
A typical workflow might start with a polynomial ring, define an ideal, compute a Groebner basis to understand it, form the affine algebra (coordinate ring) to study singularities, wrap it as a scheme to compute rational points or make a morphism to projective space, and finally use sheaf cohomology or the graded ring structure to compute genus or other numerical invariants. Toric varieties, invariant rings, differential rings, and Riemann surfaces are more specialized branches you reach for when the problem calls for their particular structure. The goal of this lesson isn't to memorize intrinsics from any one chapter — it's to recognize, when you hit a problem, which of these fifteen toolboxes it belongs to.