Skip to content

fix: Exercise 10.2.6 should bound |f x - f y| on the closed interval - #674

Open
Chessing234 wants to merge 1 commit into
teorth:mainfrom
Chessing234:ex-10-2-6-closed-interval
Open

fix: Exercise 10.2.6 should bound |f x - f y| on the closed interval#674
Chessing234 wants to merge 1 commit into
teorth:mainfrom
Chessing234:ex-10-2-6-closed-interval

Conversation

@Chessing234

Copy link
Copy Markdown
Contributor

Exercise 10.2.6 asks to show that a function which is continuous on [a,b], differentiable on (a,b), and whose derivative is bounded in absolute value by M on (a,b), satisfies |f(x) - f(y)| ≤ M|x - y| for all x, y in [a,b]. The exercise then defines Lipschitz continuity in exactly those terms, so the closed interval is the point of the statement.

The formalized version quantifies x and y over Set.Ioo a b instead, which drops precisely the two endpoints. Nothing is false as stated, but it is strictly weaker than the text, and the weakening is visible from the hypotheses: for x, y ∈ (a,b) we have [x,y] ⊆ (a,b), so hderiv alone gives Corollary 10.2.9 everything it needs and hcont is never used. Continuity on [a,b] is hypothesized precisely so that the bound extends to x or y equal to a or b, which is where the mean value theorem has to be applied on [x,y] with differentiability only on the interior.

The change is to the binders hx and hy only; the body is sorry, so no proof is affected.

I could not build locally on this machine — the Mathlib cache is not present and there is not enough free disk to fetch it — so this relies on build_book.yml in CI rather than a local lake build Analysis.Section_10_2.

Exercise 10.2.6 asks to show that a function which is continuous on
[a,b], differentiable on (a,b), and has |f'| bounded by M on (a,b) is
Lipschitz with constant M on its whole domain [a,b]; the exercise
closes by defining Lipschitz continuity in those terms.  As formalized,
x and y were quantified over the open interval (a,b) instead, which
leaves out exactly the two endpoints where the Lipschitz bound is not
immediate from the mean value theorem.

The tell is that the continuity hypothesis is unused in the weaker
version: for x, y in (a,b) the segment [x,y] is contained in (a,b), so
`hderiv` alone already supplies the continuity that Corollary 10.2.9
needs, and `hcont` never enters the argument.  It is needed only to
pass to the endpoints.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant