In this paper we examine the most commonly used algorithms for the computation of heart rate variability as well as some other interesting approaches. The aim of the paper is to study the problem from a different point of view, that of the systems designer. The selected algorithms are compared to each other through experimental analysis and theoretical study. The comparison criteria are efficiency, complexity, size of the object code, memory requirements, power consumption, parallel complexity and speedup achieved, ability to response in real time, simplicity of the interface and implementation in hardware. The motivation is strong since heart rate variability is an interesting problem which finds application not only in conventional computing systems, but also in small, even wearable devices implemented using embedded systems technology or directly in hardware. The computation of heart rate variability for a set of signal recordings and the classification achieved for these signals with each examined method are also presented. Signals have been recorded from young and elderly subjects and the examined methods are used to classify them into these two distinct groups.