Software variability can be supported by providing adaptations on top of a program's core behavior. For defining and composing adaptations in a program, different paradigms have been proposed. Two of them are feature-oriented programming and context-oriented programming. This paper compares an exemplar implementation of each paradigm. For the comparison, a common case study is used in which we detail how adaptations are defined, expressed, and composed in each paradigm. Based on the case study, we uncover similarities and differences of each implementation, and derive a set of characteristics that identify each of them. The experiment shows several overlapping similarities between the two implementations, which is an indicator that there is a similar core set of characteristics for each paradigm. This finding brings the two seemingly disjoint research directions together, and can stimulate future research both in the direction of merging features and context as well as to improve the characteristic strengths of each paradigm.