OK, so let's say that I claim that [itex]f[/itex] is continuous on some interval, [itex]I[/itex], let's say the open unit interval. Then if you give me an [itex]x_0 \in I[/itex] and an [itex]\epsilon > 0[/itex] I should be able to give you a [itex]\delta[/itex] such that [itex]|f(x_0) - f(x)| < \epsilon[/itex] whenever [itex]|x_0 - x| < \epsilon[/itex]. Now, it is important to note that I get to "see" [itex]\epsilon[/itex] AND [itex]x_0[/itex] before I have to come up with [itex]\delta[/itex].
Now, if I claim that [itex]f[/itex] is uniformly continuous on some interval, [itex]I[/itex], say, the closed unit interval, then given an [itex]\epsilon > 0[/itex] I must come up with a [itex]/delta[/itex] that will work FOR ALL [itex]x \in I[/itex]. That is, I don't get to "see" [itex]x_0[/itex] before I come up with [itex]\delta[/itex].
So, why is this important? Well, if [itex]I[/itex] is compact and [itex]f[/itex] is continuous on [itex]I[/itex] then [itex]f[/itex] is uniformly continuous on [itex]I[/itex]. Many of the theorems about derivatives, integrals, approximation of functions, etc, in analysis (at least at the undergrad level, which is all I know) require that the function be continuous on some closed and bounded (and hence compact) interval. This uniform continuity can then used to prove whatever is being proven.