A method of determining probabilities associated with a sequence of events in a system, for example an item of software or a software system being developed. A scenario, e.g. a message sequence chart, MSC (1,51), is parsed into a directed acyclic graph, DAG (21), comprising nodes (22) and edges (24), by associating the edges to messages (8,10) and sequential relations between consecutive events in the MSC (1,51). A Bayesian belief network, BBN (31), is formed using the DAG (21). BBN conditional probability tables associated with each event are computed, and the overall probabilities associated with the sequence of events is determined from the BBN conditional probaility tables. This can be used for example to provide analysis of reliability early in a soft ware development cycle. Alternative applications, e.g. to project management, are possible.
展开▼