A function $f:X\to Y$ between topological spaces is continuous if and only if $f(\overline{A})\subseteq \overline{f(A)}$ for every subset $A\subseteq X$, where the bars denote closure. For a continuous $f$, pull back an open neighbourhood of $f(x)$ for $x\in \overline{A}$; its preimage meets $A$, so the neighbourhood meets $f(A)$. Conversely, apply the inclusion to $A=f^{-1}(C)$ with $C$ closed. Then $f(\overline{A})\subseteq C$ implies $\overline{A}\subseteq A$, so the preimage of every closed subset is closed and $f$ is continuous.

A function of several variables is separately continuous when it is continuous in each variable with the others fixed. This is weaker than joint continuity in the product topology. Functions with narrow rectangular supports provide useful examples distinguishing convergence in topologies generated by separate continuity from Euclidean convergence.

For a continuous map into a metric space, $f(\overline{A})\subseteq \overline{f(A)}$: an open neighborhood of an image point pulls back to a neighborhood meeting $A$. A bounded set lies in a closed ball, and its closure lies in the same ball because distance from the centre is a continuous function. The displayed boundedness conclusion follows. Neither compactness of the domain nor uniform continuity of the map is needed.

A bounded linear operator between normed spaces is approximable if it is an operator norm limit of bounded finite-rank operators. If the target is a Banach space, every approximable operator is a compact operator. For a given tolerance, approximate the operator in norm by one finite-rank operator, cover that operator's bounded unit-ball image by finitely many small balls in its finite-dimensional range, and transfer the cover across the operator-norm error. This makes the original image totally bounded; its closure is complete in the Banach target and therefore compact. The converse on arbitrary Banach spaces would require an additional approximation hypothesis and is not part of this assertion.

An isomorphic embedding of normed vector spaces is an injective bounded linear operator whose inverse on its image is bounded. Equivalently $\Vert Tx\Vert \ge c\Vert x\Vert$ for some $c>0$. The two-sided bounds make the domain and its image carry equivalent norms. If the domain is a Banach space, a convergent sequence $T{x}_n$ has a Cauchy sequence of preimages, so its limit remains in the image; thus the image is closed.

A quotient operator is a surjective bounded linear operator that is an open map. It is equivalent to requiring a ball about zero in the image of the open unit ball. Translation and scaling then prove being an open map at every point. For a Banach space domain, a bounded linear operator is a quotient operator exactly when its Banach-space adjoint is an isomorphic embedding: a lower bound on the adjoint gives approximate ball-surjectivity by Hahn-Banach separation, and successive approximation for quotient operators gives exact preimages.

Take $X=[0,1]$ with its usual topology and $Y$ the Sierpiński space with open singleton $\{1\}$. The map $f(0)=0$, $f(x)=1$ for $x>0$ is continuous, since $f^{-1}(\{1\})=(0,1]$ is relatively open in $X$. The sets $K_n=[0,1/n]$ are nonempty, nested and compact, butThe fiber of $1$ is not closed, which is exactly where the compact-intersection proof fails. This distinguishes the valid Hausdorff-codomain theorem from a false extension to arbitrary codomains.

For a stock with constant continuous dividend yield $\delta$, holding one fixed share and depositing each dividend in a bank account at rate $r$ gives the displayed wealth process. Since $d{A}_t=r{A}_{t}dt+\delta S_{t}dt$, this is a self-financing portfolio: its only cash inflow to the bank is the dividend earned by its stock holding. A claim paying $n{X}_T$ is replicated by $n$ copies and costs $n{X}_t$, including the known accrued-dividend balance at time $t$.

A different claim paying only $n{S}_T$ costs $n{S}_t{e}^{-\delta (T-t)}$ in the Black-Scholes model with dividends. It is replicated by $n{e}^{-\delta (T-t)}$ shares with dividends reinvested in new shares: the share count grows at rate $\delta$ and reaches $n$ at expiry. Thus banking dividends and reinvesting them produce distinct terminal payoffs and distinct current values, even though both strategies are self-financing.

For smooth inequalities $c_j\ge 0$, the displayed intersection of half-spaces contains the contingent tangent cone. This follows by differentiating each active feasibility inequality along a feasible tangent sequence. The linear independence constraint qualification, and more generally the Mangasarian–Fromovitz constraint qualification, ensure equality in the inequality-only setting. Without a qualification the inclusion can be strict: for $c(x)=-x^2\ge 0$, the feasible set is $\{0\}$ and its tangent cone is $\{0\}$, while $c^{\prime }(0)=0$ makes the linearized cone the whole real line.

An operation associated with a context-free grammar production, performed when that production is recognized. Semantic actions can construct an abstract syntax tree by combining the trees attached to its child nonterminal symbols.

A context-free grammar is ambiguous when some accepted token sequence has two distinct parse trees. Giving a language an unambiguous context-free grammar requires specifying grouping in the productions themselves, rather than relying on an unstated parser conflict rule.

Store the start nonterminal symbol on a stack. An epsilon transition nondeterministically replaces its top nonterminal symbol by a production rule right-hand side. A top terminal symbol can be removed only while consuming the same input symbol. With a protected stack bottom marker, enter a final control state when that marker is exposed. Requiring all input to have been read implements acceptance by final state. The stack records the unprocessed suffix of a leftmost derivation, proving equality of the recognized and generated formal languages.

A contact line is pinned when its position remains fixed while the surrounding drop shape or contact angle changes. Surface heterogeneity or deposited material can sustain such a state over a range of angles. Evaporation from a pinned drop then requires compensating liquid motion and can generate outward solute transport associated with the coffee-ring effect. A model with a freely receding line and fixed contact angle has a different boundary constraint and need not predict the same flow direction.

A rooted tree whose leaves describe primitive regions and whose internal nodes describe Boolean operations. A half-space leaf can hold an implicit equation and an inside-sign convention. Evaluate membership from the leaves toward the root; interval arithmetic can propagate conservative inside/outside bounds over a box instead of a single point.

A constructible line is a line through two distinct constructible points. If the available coordinates lie in a real field $F$, a line through $(a,b)$ and $(c,d)$ has equation $(d-b)(x-a)-(c-a)(y-b)=0$, with coefficients in $F$ and a nonzero normal vector. Intersecting two distinct such lines introduces no new field elements, because solving a nonsingular pair of linear equations uses only field operations.

A constructible circle has a constructible point as its centre and radius equal to the distance between two constructible points. When those coordinates lie in a real field $F$, its equation $(x-a{)}^2+(y-b{)}^2=r^2$ has coefficients in $F$ because the squared radius is a sum of squared coordinate differences. Intersecting it with a constructible line reduces to a quadratic equation. Intersecting two distinct circles reduces to the same situation by subtracting their normalized equations to obtain the radical axis. Thus each isolated real intersection requires at most a real quadratic field extension.

The hypergraph whose vertices are the variables of a constraint satisfaction problem and whose hypergraph edges are its constraint scopes. Several constraints can have the same scope. An incidence drawing uses extra scope nodes for display; these do not become extra variables of the original problem. Ordinary treewidth of a constraint hypergraph refers to treewidth of its primal graph, equivalently to a tree decomposition covering each whole scope, not to generalized hypergraph width measures.

A collection of propositional formulas is consistent if it does not derive a contradiction. Equivalently, in a standard classical calculus, it does not derive both a logical formula and its negation.

For a nontrivial connected graph, return the internal vertices of a rooted tree in a depth-first search tree. Undirected depth-first search has only ancestor-descendant edges, so every edge meets an internal vertex; the internal vertices are connected along parent paths. For a maximum-size proper matching in a rooted tree $M$ of that tree, existence of which is assumed, $|I|=|V|-|L|\le |V|-|U(M)|=2|M|$. Every vertex cover has size at least $|M|$, proving the factor two. The implementation needs only depth-first search, hence linear time complexity; it does not compute the analysis matching in a graph.

The torus and Klein bottle are not homeomorphic because only the former is orientable. Adding a projective plane to either by connected sum of surfaces makes a nonorientable surface with Euler characteristic minus one. The classification of compact surfaces identifies both with three projective planes. Thus connected-sum cancellation can fail even for closed connected surfaces.