I think the specific reference you are referring to is
All the methods mentioned above are convergent. In fact, convergence is a condition sine qua non for any numerical scheme.
This section is referring to numerical methods to solve a differential equations by "iterative" methods- finding one function after another hopefully getting closer and closer to a true solution. That is saying that the sequence of functions
converges to a true solution. It might be that a sequence of functions given by a particular method converges very slowly to a true solution (so it's not a very useful method) or that, no matter how many terms you take it is still slightly off a true solutions (taking higher and higher terms in Fourier series
around a point at which the function is not continuous will always "miss" the function in some neighborhood of that point- the neighborhood gets smaller, but the "error" never goes to 0. See "Gibbs phenomenon")
However, whether a sequence of functions converges slowly or slightly misses a true solution, it might still be a useful method for some purposes. If, on the other hand, the sequence does not converge at all, it can't possibly be a useful method for finding a solution, even approximately! That is why convergence is a "sine qua non" ("without which, not") to even consider an interation method. Without at least convergence, it can not be a good method!