According to one aspect of the present disclosure, a computer-implemented method is disclosed in which a software development log whose entries describe potential features for a future release of a first software program is accessed to obtain a text description of a feature from an entry. The text description is decomposed using a predefined n-gram decomposition technique to produce a first set of n-grams. The method includes determining whether a second software program includes the feature. The determining includes comparing the first set of n-grams to a second set of n-grams which describe the second software program, and calculating a similarity metric based on the comparing that indicates an extent to which matches for the n-grams in the first set are present in the second set. The software development log is updated based on the determining.