We prove uniform continuity by contradiction. If it fails, there is such that, for every , we can choose with
By sequential compactness, pass to a subsequence with . The triangle inequality gives
By continuity, both and converge to . A second triangle inequality therefore gives
contradicting the lower bound. Thus the requested , independent of , exists for every .
Both properties are preserved by a continuous map.
For sequential compactness, take any sequence in and choose with . If is sequentially compact, some subsequence satisfies . By continuity, . Hence is a sequentially compact space.
For connectedness, suppose instead that is a separation into disjoint nonempty relatively open sets. Then and are disjoint nonempty open sets of whose union is , contradicting connectedness. Therefore is a connected space.
First verify that is a vector space. Sums and scalar multiples preserve continuity and the condition at zero. For every partition of an interval , the triangle inequality and absolute homogeneity of a norm of the absolute value imply
Taking suprema yields
For the second equality follows directly from . In particular, finite variation is preserved, so is closed under these operations.
Nonnegativity is immediate. For any , use the partition of an interval , omitting repeated endpoints if necessary, to get
Thus forces , and conversely has zero total variation of a function. These observations prove that is a norm on .
Applying the same bound to gives
Therefore convergence in the anchored variation norm implies uniform convergence on . The anchoring condition removes the nonzero constant functions which otherwise have zero total variation of a function.
A norm on a real vector space is a map such that, for and ,
The last property is the triangle inequality; the middle property is absolute homogeneity of a norm. The induced metric is
Positivity and separation follow from the norm axioms, symmetry from absolute homogeneity of a norm with , and the triangle inequality from .
For , the triangle inequality gives
Taking the supremum proves . Hence convergence for implies convergence for .
The converse is false. For , take and . Its supremum norm is , whereas the binomial theorem gives
The alternating coefficients cancel on ; the polynomial coefficient sum norm measures their absolute values before any cancellation.
The space of polynomials is not complete for the supremum norm metric . Consider
The geometric series has sum , and for its tail satisfies
Thus converges uniformly to and is a Cauchy sequence for .
There is no polynomial agreeing with on : otherwise the polynomial would vanish at infinitely many points and therefore be the zero polynomial. Evaluating this identity at gives . By uniqueness of the uniform limit, has no limit in . This proves the claimed failure of completeness.
Extend the coefficient lists of each polynomial by zeros. Then is finite because only finitely many terms are nonzero. It is nonnegative and symmetric. Moreover, holds exactly when every coefficient agrees, which by uniqueness of polynomial coefficients is equivalent to .
If , the triangle inequality for the absolute value gives
Thus satisfies all the metric axioms. It is the metric induced by the polynomial coefficient sum norm.
The arctangent pullback metric is not complete. Take . Since , its transformed sequence is Cauchy, and hence is a Cauchy sequence for . If it converged in to , then
would force , which is impossible. The missing endpoint of explains the failure of completeness, even though part (i) shows that the metric topologies agree.
The inverse tangent is a homeomorphism from the real line with its Euclidean metric onto with its subspace topology: and its inverse are continuous. It is also an isometry from onto , because . Consequently, for ,
For a direct open ball interpretation, is -open by continuity of . Conversely, continuity of at supplies a sufficiently small -ball inside any prescribed -ball about . Thus both metrics generate exactly the same open sets.
A metric space is a complete metric space if every Cauchy sequence in converges to a point of . Explicitly, the hypothesis is
and the conclusion is that some satisfies .
It is a connected space if there are no disjoint nonempty open subsets of with . Here openness refers to the metric topology of .
It is a path-connected space if, for every , there is a continuous map with and . Such a map is a path. Every path-connected space is a connected space: a separation would pull back along a path joining its two sides to a separation of the interval .
Let . Between the two middle order statistics, and the slope is zero. To the left decreases and to the right it increases. Consequently all joint maximum-likelihood estimates are
where the constant minimum is
The strict positivity follows from distinct observations. Thus the location maximum-likelihood estimate is not unique, while the rate maximum-likelihood estimate is unique and the same for every location in the displayed interval. Choosing the arithmetic midpoint of the two middle values is the usual sample median convention and gives one member of this entire maximizing family.
For each fixed , maximizing the likelihood function over is equivalent to minimizing . Order the distinct observations as and write . Between successive order statistics, with exactly observations to the left,
Thus decreases up to and increases after it. Its unique minimum is at the sample median:
At a fixed with , the log-likelihood function has derivative and second derivative , so its unique maximizing rate is . The maximized likelihood function is , which increases when decreases. Hence the joint maximum-likelihood estimates are
This is Laplace location and scale maximum likelihood, expressed with the reciprocal scale.
A finite rate maximum-likelihood estimate requires a positive total absolute deviation. For distinct observations and odd , that condition holds. For , however, gives and as . There is no finite joint maximum-likelihood estimate in the stated parameter space; the formula with zero denominator is not an admissible estimate.
The joint maximum-likelihood estimators are functions of the observations selecting a maximizing parameter pair:
The estimates after observation maximize the same likelihood function with . Maximization may be nonunique, as for an even sample size, or may fail to yield a finite admissible pair, as in the one-observation case below. The definition does not turn a boundary supremum into an attained maximum.
By independence, the likelihood function is the product of the individual probability density functions:
Writing , the log-likelihood function is . Here is a rate, the reciprocal of the scale in the usual Laplace distribution notation.
For this rate parametrization of the Laplace distribution, is its mean and median. By symmetry, , while has exponential distribution with rate , so
The integral follows by two applications of integration by parts. The central limit theorem therefore gives
With known and the supplied cutoff 2, an approximate 95% confidence interval is
This known-rate Laplace location confidence interval uses the sample mean; the later maximum-likelihood estimator of location need not be used to construct this valid asymptotic confidence interval.

There are unlisted articles, also show them or only show them.