As two different research topics with much overlap, dependability and security of computer/communication systems have respective long and rich history. The development of the techniques for their modeling and analysis thus have followed distinct but convergent paths. In essence, diverse attributes and the fundamental difference between the nature of the failures bring in different concerns for dependability and security analysis during their modeling process. Taking the understanding of the basic concepts/attributes as a point of departure, this paper intend to carry out a comparative study on the analytical models of computer system dependability and security. Also, by examining the state-of-the-art quantitative techniques and sound modeling methodologies for dependability evaluation, e.g., combinatorial and stochastic methods, we attempt to explore why and how those methods can be extended to evaluate computer system security. Furthermore, we take our developed autonomic detection coordinator (for intrusion detection) as a case study to conduct the comparative analysis.