Subdirect Products of Rings

Q: Is being a subdirect product a property of the ring or of the map?

Of the map. A ring is a subdirect product *of a given family, via a given embedding*. The same ring usually has many inequivalent representations, and a theorem such as (12.5) asserts the existence of one with factors of a prescribed kind, not its uniqueness.

Q: If every factor is a field, must the ring be a field?

No. Z is a subdirect product of the fields F_p. Being a field is not closed under products — F_2 × F_3 is not a field — so the transfer proposition does not apply. What does transfer is commutativity and reducedness, and indeed Z is a commutative reduced ring.

Q: How does a subdirect product differ from a fibre product?

A fibre product A ×_C B of two surjections A → C B is always a subdirect product of A and B, and it is the general shape of a two-factor subdirect product: given a subdirect R ⊆ A × B one can often recover C as a common quotient. For infinitely many factors there is no comparably tidy description.

Q: Can the intersection of the kernels be zero with only finitely many factors?

Certainly — Z/6 ≅ F_2 × F_3 uses two, and the fibre product Z ×_F_2 Z uses two while failing to be a direct product. Infinitely many factors are needed only when the ring has no finite separating family, as happens for Z with prime quotients.

Executive Summary

A subdirect product representation writes a ring $R$ as a subring of a product $\prod_{i \in I} R_{i}$ in such a way that nothing is lost (the map is injective) and no factor is wasted (every coordinate map is onto). The factors are automatically quotients of $R$ , so they are simpler; the price is that $R$ is recovered only as a subring of the product, not as the whole product.

The entire notion is equivalent to a piece of ideal theory: a subdirect representation of $R$ is a family of ideals $𝔄_{i} ⊴ R$ with $⋂_{i} 𝔄_{i} = 0$ . Every later theorem in this stream — semiprime rings from prime rings, semiprimitive rings from primitive rings, reduced rings from domains — is that dictionary applied to a well-chosen family of ideals.

$⋂_{i} 𝔄_{i} = 0$ The whole criterion

OntoEvery coordinate map

(12.1)Lam's numbering

1944Birkhoff's paper

Overview

A direct product $\prod_{i \in I} R_{i}$ is easy to understand but rare: most rings do not decompose. What is common is that a ring admits many surjections onto simpler rings, and that these surjections jointly detect every nonzero element. That weaker situation is what a subdirect product records.

ε : R ↪ \prod_{i \in I} R_{i}, π_{i} \circ ε : R ↠ R_{i} for every i \in I

(12.1)

Injective on the left, surjective on each coordinate. Neither condition alone is enough.

Dropping injectivity gives nothing; dropping surjectivity of the coordinate maps gives merely a subring of a product, which carries no information at all, since every ring is a subring of some product. The two conditions together are what make the factors usable.

The notion originates in universal algebra rather than ring theory: Birkhoff introduced subdirect unions for arbitrary algebraic systems, and the ring-theoretic use — decomposing a ring into building blocks that cannot be decomposed further — is a specialisation. The building blocks are treated in Subdirectly Irreducible Rings and the Little Ideal.

Learning Objectives

State $(12.1)$ precisely, including what makes a representation trivial.
Prove the equivalence between subdirect representations and separating families of ideals.
Recognise when a subring of $\prod_{i} R_{i}$ is a subdirect product and when it is not.
Determine which properties pass from the factors $R_{i}$ back to $R$ , and which do not.
Write $ℤ$ as a nontrivial subdirect product of finite fields.
Build the fibre product $ℤ \times_{𝔽_{2}} ℤ$ and verify it is subdirect but indecomposable.

Definitions

Definition(12.1)Subdirect product representation

Let $R$ and ${R_{i} : i \in I}$ be rings and let $ε : R \to \prod_{i \in I} R_{i}$ be an injective ring homomorphism. We say $ε$ *represents $R$ as a subdirect product of the $R_{i}$ * if for every $i \in I$ the composite $π_{i} \circ ε : R \to R_{i}$ with the $i$ -th coordinate projection is surjective. Informally one says $R$ is a subdirect product of the $R_{i}$ .

The representation is called trivial if some coordinate map $π_{i} \circ ε$ is an isomorphism. In that case the single factor $R_{i}$ already recovers $R$ and the remaining factors carry no extra information.

$\prod_{i \in I} R_{i}$: The direct product: all families $(r_{i})_{i \in I}$ with $r_{i} \in R_{i}$ , with componentwise operations and identity $(1)_{i \in I}$ .
Coordinate map: $ϕ_{i} : = π_{i} \circ ε : R \to R_{i}$ . Its kernel $𝔄_{i}$ is an ideal of $R$ and $R / 𝔄_{i} ≅ R_{i}$ when $ϕ_{i}$ is onto.
Separating family: A family of ideals ${𝔄_{i}}$ of $R$ with $⋂_{i} 𝔄_{i} = 0$ ; equivalently, the family of quotient maps jointly detects every nonzero element.
Direct representation: The special case in which $ε$ is also surjective, so $R ≅ \prod_{i} R_{i}$ . Every direct representation is subdirect; the converse fails badly.
Irredundant representation: One in which no proper subfamily of ${𝔄_{i}}$ still has zero intersection. Not required by $(12.1)$ and not always achievable.

All rings have an identity, homomorphisms preserve it, and ideal without qualification means two-sided ideal.

Core Concepts

From maps to ideals and back

An element of $R$ is invisible to the representation exactly when it dies under every coordinate map. Injectivity of $ε$ therefore says precisely that the kernels intersect in zero, and surjectivity of the coordinate maps says the factors are quotients rather than arbitrary overrings. Passing between the two pictures is mechanical, and it is worth doing once explicitly.

Family

{𝔄_{i}}

with

⋂ 𝔄_{i} = 0

Quotient maps

R ↠ R / 𝔄_{i}

Embedding

R ↪ \prod_{i} R / 𝔄_{i}

What triviality means

A representation is trivial exactly when some kernel $𝔄_{i}$ is zero. Such a representation is legitimate but useless: it decomposes $R$ into $R$ itself plus decoration. Rings for which every subdirect representation is trivial are the atoms of the theory; they are exactly the rings possessing a smallest nonzero ideal.

How far the image sits from the product

The gap between $ε (R)$ and $\prod_{i} R_{i}$ can be enormous. For $ℤ ↪ \prod_{p} 𝔽_{p}$ over all primes, the source is countable and the target has the cardinality of the continuum, yet every coordinate map is onto. Surjectivity coordinate by coordinate is a far weaker demand than surjectivity onto the product.

Key Results

Proposition(12.1)′Ideal-theoretic criterion

Let $R$ and ${R_{i} : i \in I}$ be rings. Then $R$ can be represented as a subdirect product of the $R_{i}$ iff for each $i$ there is a surjective ring homomorphism $ϕ_{i} : R \to R_{i}$ such that $⋂_{i \in I} \ker ϕ_{i} = 0$ . Moreover the resulting representation is trivial iff $\ker ϕ_{i} = 0$ for some $i$ .

Proof

Suppose $ε : R \to \prod_{i} R_{i}$ is a subdirect representation and put $ϕ_{i} = π_{i} \circ ε$ , which is onto by hypothesis. An element $r$ lies in every $\ker ϕ_{i}$ iff every coordinate of $ε (r)$ vanishes, i.e. iff $ε (r) = 0$ ; since $ε$ is injective this forces $r = 0$ , so $⋂_{i} \ker ϕ_{i} = 0$ .

Conversely, given surjections $ϕ_{i}$ with $⋂_{i} \ker ϕ_{i} = 0$ , define $ε (r) = (ϕ_{i} (r))_{i \in I}$ . This is a ring homomorphism because the operations on the product are componentwise, and it sends $1$ to $(1)_{i}$ . Its kernel is $⋂_{i} \ker ϕ_{i} = 0$ , so $ε$ is injective, and $π_{i} \circ ε = ϕ_{i}$ is onto for each $i$ . Hence $ε$ is a subdirect representation.

For the last claim, a surjection is an isomorphism precisely when its kernel vanishes, so some coordinate map is an isomorphism iff some $\ker ϕ_{i} = 0$ .

Corollary(12.1)″The dictionary

Up to isomorphism of the factors, the subdirect representations of $R$ are exactly the families of ideals ${𝔄_{i} : i \in I}$ of $R$ with $⋂_{i} 𝔄_{i} = 0$ , the associated factors being $R_{i} = R / 𝔄_{i}$ . A nonzero ring $R$ therefore admits a nontrivial subdirect representation iff the intersection of its nonzero ideals is zero.

ε : R ↪ \prod_{i \in I} R / 𝔄_{i}, r \mapsto (r + 𝔄_{i})_{i \in I}, ⋂_{i \in I} 𝔄_{i} = 0

(12.1)″

The canonical form of a subdirect representation. Every representation is isomorphic to one of this shape.

Proposition—Transfer of properties

Let $R$ be a subdirect product of ${R_{i} : i \in I}$ and let $𝒫$ be a class of rings.

If $𝒫$ is closed under homomorphic images and $R \in 𝒫$ , then every $R_{i} \in 𝒫$ .
If $𝒫$ is closed under subrings and arbitrary direct products and every $R_{i} \in 𝒫$ , then $R \in 𝒫$ .

Proof

(1) Each $R_{i}$ is the image of the surjection $π_{i} \circ ε$ , hence a homomorphic image of $R$ . (2) The product $\prod_{i} R_{i}$ lies in $𝒫$ by closure under products, and $ε$ identifies $R$ with a subring of it, so $R \in 𝒫$ by closure under subrings.

Both halves are used constantly. Commutativity, reducedness, and satisfaction of a fixed polynomial identity are closed under subrings, products and images, so they transfer in both directions — this is exactly what makes the commutativity theorems of §12 provable by reduction. Being a domain, being noetherian, and being artinian are closed under none of the three combinations required, and do not transfer.

Remark—Why nontriviality matters

Every ring $R$ is a subdirect product of the one-element family ${R}$ . A theorem asserting that some class of rings consists of subdirect products of nicer rings is therefore only informative when the factors are constrained — prime, left primitive, domain, subdirectly irreducible. The constraint, not the existence of a representation, is the content.

Proof Techniques and Method

How these arguments are actually assembled, and which step to reuse.

Building a subdirect representation is a three-move routine, and every theorem in this section follows it.

Move 1

Name the ideal family

Choose the ideals whose quotients you want as factors: all prime ideals, all left primitive ideals, all minimal primes, or the ideals $𝔪_{a}$ maximal with respect to excluding a fixed $a \neq 0$ .

Move 2

Prove the intersection is zero

This is always where the mathematics lives. It is usually a radical computation: $⋂ 𝔭 = {Nil}_{*} R$ , or $⋂ (primitive) = rad R$ , and the hypothesis on $R$ says that radical vanishes.

Move 3

Read off the factors

Each $R / 𝔄_{i}$ automatically belongs to the class defined by the ideals chosen — $R / 𝔭$ is prime, $R / 𝔪_{a}$ is subdirectly irreducible, and so on.

The converse direction is even shorter: given a representation with factors in a class closed under subrings and products, the ring inherits the class membership. Almost every iff statement in this section is Move 2 in one direction and this one-line closure argument in the other.

Worked Example

The integers as a subdirect product of finite fields

Take $𝔄_{p} = p ℤ$ for every prime $p$ . Each quotient $ℤ / p ℤ = 𝔽_{p}$ is a field, and an integer divisible by every prime is zero, so $⋂_{p} p ℤ = 0$ .

ε : ℤ ↪ \prod_{p prime} 𝔽_{p}, n \mapsto (n mod p)_{p}

(E.1)

Nontrivial: every kernel $p ℤ$ is nonzero, so no coordinate map is injective.

Any infinite set of primes already works, and so does the family ${(p_{i}^{n_{i}})}$ for distinct primes $p_{i}$ and any exponents $n_{i} \geq 1$ — the ring $ℤ$ has uncountably many genuinely different subdirect representations. The image is minuscule: $ε (ℤ)$ is countable while the product is not.

A subdirect product that is not a direct product

Let $S = {(a, b) \in ℤ \times ℤ : a \equiv b (\mod 2)}$ . It contains $(1, 1)$ and is closed under subtraction and multiplication, since $a \equiv b$ and $c \equiv d$ modulo $2$ give $a c \equiv b d$ . So $S$ is a subring of $ℤ \times ℤ$ of index $2$ .

Coordinate maps are onto. Given $a \in ℤ$ , the element $(a, a)$ lies in $S$ and projects to $a$ in either coordinate.
Kernels. $\ker π_{1} |_{S} = {(0, b) : b even} = 0 \times 2 ℤ$ and $\ker π_{2} |_{S} = 2 ℤ \times 0$ ; both are nonzero and they meet in $0$ .
Conclusion. $S$ is a nontrivial subdirect product of $ℤ$ and $ℤ$ , and $S \neq ℤ \times ℤ$ since $(1, 0) \notin S$ .

$S$ is not a direct product of two rings at all. Its idempotents are the pairs $(e, f)$ with $e, f \in {0, 1}$ and $e \equiv f (\mod 2)$ , namely $(0, 0)$ and $(1, 1)$ ; an indecomposable ring admits no nontrivial direct decomposition. So $S$ is genuinely a subdirect and not a direct product — the two notions are separated by a single example.

Comparison and Classification

Four ways a ring can sit inside a product
Situation	$ε$ injective	Coordinate maps onto	$ε$ onto	Information content
Arbitrary subring of $\prod R_{i}$	yes	not required	no	none — every ring is one
Subdirect product	yes	yes	no	factors are quotients of $R$
Trivial subdirect product	yes	yes	no	some factor already equals $R$
Direct product decomposition	yes	yes	yes	$R$ splits by central idempotents

Does the property pass in the indicated direction?
	$R \Rightarrow$ each $R_{i}$	each $R_{i} \Rightarrow R$
Commutative	yes	yes
Reduced	yes	yes
Satisfies a fixed polynomial identity	yes	yes
Semiprime	no	yes
Domain	no	no
Left noetherian	yes	no
Left artinian	yes	no
Simple	no	no

Does the property pass in the indicated direction?

The semiprime row is the pattern the whole section exploits: the property fails to descend to the factors as stated but is created by them, because prime factors force semiprimeness upstairs. Read the row for domain as a warning: a subdirect product of domains is reduced but almost never a domain, as $ℤ ↪ \prod_{p} 𝔽_{p}$ shows in one direction and $ℤ \times ℤ$ in the other.

Relationship Map

The three main theorems of §12 are one construction applied to three different ideal families. The pattern is worth memorising as a single picture.

Subdirect representation of $R$ — choose a family of ideals with zero intersection
- All prime ideals
  - intersection is ${Nil}_{*} R$
  - zero iff $R$ is semiprime
  - factors are prime rings
- All left primitive ideals
  - intersection is $rad R$ by $(11.5)$
  - zero iff $R$ is semiprimitive
  - factors are left primitive rings
- All minimal prime ideals
  - intersection is again ${Nil}_{*} R$
  - zero and $R$ reduced gives completely prime factors
  - factors are domains
- Ideals maximal without a fixed $a \neq 0$
  - intersection is zero for trivial reasons
  - no hypothesis on $R$ needed
  - factors are subdirectly irreducible

Subrings of $\prod_{i} R_{i}$ no constraint at all

Subdirect productsevery coordinate map onto

Direct products

ε

onto as well

Trivial representationssome coordinate map an isomorphism

The innermost band is not contained in the one above it in general: a trivial representation need not be direct. The picture records typical containments, not a chain of implications.

Applications and Industry Use

Applications here means where this structure is used — inside mathematics and in the engineering and computing disciplines that consume it.

Computer arithmetic

Residue number systems

Representing an integer by its residues modulo pairwise coprime moduli is the map $ℤ \to \prod_{i} ℤ / m_{i}$ . Addition and multiplication become componentwise and carry-free, which is why RNS arithmetic appears in DSP hardware and in RSA implementations.

Symbolic computation

Multi-modular algorithms

Computer algebra systems compute a result over $ℤ$ by computing it modulo many primes and reconstructing. Correctness rests on exactly the injectivity in $ℤ ↪ \prod_{p} 𝔽_{p}$ ; the cost model rests on the factors being small.

Boolean algebra

Stone representation

A Boolean ring is commutative and reduced, and all its quotients are Boolean, so its subdirectly irreducible quotients are Boolean fields, i.e. $𝔽_{2}$ . Every Boolean ring is therefore a subdirect product of copies of $𝔽_{2}$ — a ring of ${0, 1}$ -valued functions.

Ring theory proper

Reduction machinery

The honest main use is internal: subdirect representations let a theorem be checked on primitive or prime rings and then exported to semiprimitive or semiprime rings, as in The Jacobson and Herstein Commutativity Theorems.

Two of these are not really analogies. The Chinese Remainder Theorem is the statement that $ℤ / m_{1} \dots m_{r} \to \prod_{i} ℤ / m_{i}$ is a direct — not merely subdirect — representation when the $m_{i}$ are pairwise coprime, and the residue number systems used in hardware are that theorem implemented in silicon.

Design Considerations

Design considerations here means the choices made when modelling a problem with these algebraic structures.

Which ideal family? Small factors mean many of them and an image that is hard to describe; large factors mean few of them but little simplification. All primes gives prime factors; minimal primes gives fewer factors and, in the reduced case, domains.
Irredundancy is optional. $(12.1)$ does not demand that the family be minimal, and for infinite families a minimal separating family may not exist. Do not build proofs that assume one.
Do not expect the image to be describable. Identifying $ε (R)$ inside $\prod_{i} R_{i}$ is usually harder than the original problem, and is almost never needed: injectivity plus surjectivity of the coordinate maps is what proofs consume.
Check closure before transferring. Before concluding a property of $R$ from a property of the factors, verify closure under subrings and under arbitrary products; the second condition eliminates all chain conditions.
Two-sided ideals only. The kernels here are two-sided, so one-sided phenomena — left primitivity as against right primitivity, one-sided chain conditions — are not visible to the construction and must be handled separately.

Standards and Notation

Standards here covers notation, symbol and markup standards, and reference implementations, rather than material or design codes.

Lam's notation

ε : R \to \prod_{i \in I} R_{i}

, with the representation named by the map

Universal algebraBirkhoff's term is subdirect union; subdirect product is now standard in both fields

Common alternativeWriting

R \leq_{s d} \prod_{i} R_{i}

or calling

R

a subdirect sum, chiefly in older literature

Products

\prod

for direct product,

⨁

for the direct sum; these differ for infinite

I

and the direct sum is not a ring with identity

MarkupPresentation MathML per ISO/IEC 40314; operator symbols per ISO 80000-2

ImplementationsGAP and Sage build

\prod_{i} R_{i}

directly; subdirect images are constructed as the image of an explicit tuple of quotient maps

Failure Modes and Common Mistakes

Do not conclude that the factors are unique. They are determined by the chosen ideal family and nothing else, and different families can give non-isomorphic factor sets for the same ring.
Do not forget the surjectivity requirement. Without it the statement every ring is a subdirect product of fields would be vacuously false-flavoured nonsense rather than a genuine restriction.
Do not assume a subdirect product of domains is a domain — it is exactly a reduced ring, which is the content of Reduced Rings as Subdirect Products of Domains.
Do not treat a trivial representation as an error. It is a legitimate representation; the theorems that matter assert the existence of a nontrivial one, or the impossibility of any.

Historical Notes and Lessons Learned

1936Stone's representation theoremStone represents an arbitrary Boolean algebra as an algebra of sets, in effect exhibiting a Boolean ring as a subdirect product of copies of the two-element field.
1944Birkhoff's subdirect unionsBirkhoff isolates the notion for arbitrary algebraic systems and proves that every algebra is a subdirect product of subdirectly irreducible ones — the universal-algebra ancestor of $(12.3)$ .
1945McCoy brings it to ringsMcCoy studies subdirectly irreducible commutative rings and uses subdirect decompositions systematically in radical theory; the prime radical is named after him and Baer.
1945–1956Radicals and semisimple classesWith Jacobson's radical available, the pattern radical zero iff subdirect product of the corresponding irreducible rings becomes the standard organising principle of general radical theory.
1960sKurosh–Amitsur formalismRadical classes are axiomatised, and the closure of a semisimple class under subdirect products is taken as one of the defining conditions, making $(12.5)$ an instance of a general theorem.

The methodological lesson is that the useful decomposition theorem was not the one that splits a ring into pieces, but the one that embeds it into a product of pieces. Insisting on genuine direct decompositions confines you to rings with plenty of central idempotents; accepting an embedding costs almost nothing and applies to every ring.

Quick Reference

Definition

ε : R ↪ \prod_{i} R_{i}

injective, every

π_{i} ε

onto

Ideal formideals

𝔄_{i}

with

⋂_{i} 𝔄_{i} = 0

, factors

R / 𝔄_{i}

Trivialsome

𝔄_{i} = 0

Always availablethe one-factor family

{R}

; hence existence alone says nothing

Transfers downany property closed under quotients

Transfers upany property closed under subrings and products

Never transfers upchain conditions, simplicity, being a domain

ReferenceLam

(12.1)

, §12

The four §12 decomposition theorems at a glance
Hypothesis on $R$	Ideal family used	Factors	Result
$R \neq 0$ , no other hypothesis	$𝔪_{a}$ maximal excluding $a$	subdirectly irreducible rings	(12.3)
$R$ semiprime	all prime ideals	prime rings	(12.5)
$R$ semiprimitive	all left primitive ideals	left primitive rings	(12.5)
$R$ reduced	all minimal prime ideals	domains	(12.7)

Frequently Asked Questions

Is being a subdirect product a property of the ring or of the map?

Of the map. A ring is a subdirect product of a given family, via a given embedding. The same ring usually has many inequivalent representations, and a theorem such as $(12.5)$ asserts the existence of one with factors of a prescribed kind, not its uniqueness.

Why must the coordinate maps be surjective?

Without surjectivity the notion is empty: every ring embeds in some product, for instance the one-factor product containing itself. Surjectivity forces each factor to be a quotient $R / 𝔄_{i}$ , so the factors are genuinely simpler objects built from $R$ rather than unrelated rings that merely happen to contain it.

If every factor is a field, must the ring be a field?

No. $ℤ$ is a subdirect product of the fields $𝔽_{p}$ . Being a field is not closed under products — $𝔽_{2} \times 𝔽_{3}$ is not a field — so the transfer proposition does not apply. What does transfer is commutativity and reducedness, and indeed $ℤ$ is a commutative reduced ring.

How does a subdirect product differ from a fibre product?

A fibre product $A \times_{C} B$ of two surjections $A \to C \leftarrow B$ is always a subdirect product of $A$ and $B$ , and it is the general shape of a two-factor subdirect product: given a subdirect $R \subseteq A \times B$ one can often recover $C$ as a common quotient. For infinitely many factors there is no comparably tidy description.

Does the theory need an identity element?

The definition does not, and universal algebra treats rings without identity as easily as rings with. This collection assumes an identity throughout, which is why the target of a representation must be the direct product and never the direct sum: the image of $1$ has a nonzero entry in every coordinate.

Can the intersection of the kernels be zero with only finitely many factors?

Certainly — $ℤ / 6 ≅ 𝔽_{2} \times 𝔽_{3}$ uses two, and the fibre product $ℤ \times_{𝔽_{2}} ℤ$ uses two while failing to be a direct product. Infinitely many factors are needed only when the ring has no finite separating family, as happens for $ℤ$ with prime quotients.

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §12 (pp. 203–204).
G. Birkhoff, “Subdirect unions in universal algebra”, Bulletin of the American Mathematical Society 50 (1944), 764–768.
N. H. McCoy, The Theory of Rings, Macmillan, 1964, chapters on subdirect sums and radicals.
N. Jacobson, Structure of Rings, American Mathematical Society Colloquium Publications 37, revised edition, 1964.
S. Burris and H. P. Sankappanavar, A Course in Universal Algebra, Graduate Texts in Mathematics 78, Springer-Verlag, 1981, Chapter II (subdirect representations).
L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988.

AI Suggested Questions

Give an example of a ring with two subdirect representations whose factor sets share no isomorphism class.
For which rings does a finite separating family of prime ideals exist, and what does that say about the ring?
How is the closure of a semisimple class under subdirect products used in the Kurosh–Amitsur axiomatisation of radicals?
Work out the image of $ℤ ↪ \prod_{p} 𝔽_{p}$ and relate it to the profinite completion of $ℤ$ .
Which categorical limit, if any, does a general subdirect product represent?
Show that a commutative von Neumann regular ring is a subdirect product of fields, and identify when the representation can be taken to be direct.
How do residue number systems choose their moduli, and what is the arithmetic cost of the reconstruction step?

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

From maps to ideals and back

What triviality means

How far the image sits from the product

06Key Results

07Proof Techniques and Method

Name the ideal family

Prove the intersection is zero

Read off the factors

08Worked Example

The integers as a subdirect product of finite fields

A subdirect product that is not a direct product

09Comparison and Classification

10Relationship Map

11Applications and Industry Use

Residue number systems

Multi-modular algorithms

Stone representation

Reduction machinery

12Design Considerations

13Standards and Notation

14Failure Modes and Common Mistakes

15Historical Notes and Lessons Learned

16Quick Reference

17Frequently Asked Questions

18Related KEVOS Topics

19References

20AI Suggested Questions

Continue learning

Executive Summary

Overview

Learning Objectives

Definitions

Core Concepts

Key Results

Proof Techniques and Method

Worked Example

Comparison and Classification

Relationship Map

Applications and Industry Use

Design Considerations

Standards and Notation

Failure Modes and Common Mistakes

Historical Notes and Lessons Learned

Quick Reference

Frequently Asked Questions

Related KEVOS Topics

References

AI Suggested Questions