In Wikipedia and PlanetMath product of uniform spaces is defined as the weakest uniformity on the Cartesian product making all the projection maps uniformly continuous.

But Springer's encyclopedia has a different (supposedly equivalent) definition. The sad part is that I don't understand their notation.

Could you explain me how to understand this (Springer's) definition of product of uniform spaces?


The notation in the Springer definition is pretty appalling:

The product of uniform spaces $(X_t,\mathfrak{A}_t),\,t\in T$, is the uniform space $(\prod X_t,\prod\mathfrak{A}_t)$, where $\prod\mathfrak{A}_t$ is the uniformity on $\prod X_t$ with as base for the entourages sets of the form $$\Big\{\big(\{x_t\},\{t_t\}\big):(x_{t_i},y_{t_i})\in V_{t_i},i=1,\dots,n\Big\},t_i\in T,V_{t_i}\in\mathfrak{A}_{t_i},n=1,2,\dots\;.$$

Here’s what I consider a somewhat more comprehensible version of this.

Suppose that $\langle X_\alpha,\mathscr{U}_\alpha\rangle$ is a uniform space for each $\alpha\in A$, where each $\mathscr{U}_\alpha$ is a diagonal uniformity. Let $X=\prod\limits_{\alpha\in A}X_\alpha$; points of $X$ have the form $x=\langle x_\alpha:\alpha\in A\rangle$. For each finite, non-empty $F\subseteq A$ and each function $U:F\to\bigcup\limits_{\alpha\in F}\mathscr{U}_\alpha$ such that $U(\alpha)\in\mathscr{U}_\alpha$ for each $\alpha\in F$ let

$$V_{F,U}=\{\langle x,y\rangle\in X\times X:\langle x_\alpha,y_\alpha\rangle\in U(\alpha)\text{ for all }\alpha\in F\}\;,$$

and let $\mathscr{V}$ be the set of all such entourages $V_{F,U}$ of the diagonal in $X\times X$. Then $\mathscr{V}$ is a base for a diagonal uniformity $\mathscr{U}$ on $X$, which we call the product uniformity on $X$.

This is analogous to the usual definition of the product topology: that defines a typical basic open set by picking a finite set $F$ of indices and an open set $U(\alpha)$ in $X_\alpha$ for each $\alpha\in F$ and letting $$B(F,U)=\{x\in X:x_\alpha\in U(\alpha)\text{ for each }\alpha\in F\}\;.$$

  • $\begingroup$ What is "diagonal uniformity"? The link it points to does not contains words "diagonal uniformity" $\endgroup$ – porton May 23 '12 at 16:46
  • $\begingroup$ @porton: A diagonal uniformity is a uniformity defined by entourages, rather than by uniform covers or pseudometrics. $\endgroup$ – Brian M. Scott May 23 '12 at 19:05
  • $\begingroup$ Can this be re-formulated in terms of Frechet filters? I spend a few minuted trying to simplify this formula into a statement about Frechet filters without success and as such deem it impossible. But may it be nevertheless possible? $\endgroup$ – porton Jun 4 '12 at 16:18
  • $\begingroup$ @porton: It seems unlikely. $\endgroup$ – Brian M. Scott Jun 4 '12 at 20:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.