That is how it is implemented, but the core idea is that a lot of problems can be broken down into subproblems that require shared, repetitive computations to solve, and the table is a way to cache/share those results for each subproblem.