This paper presents PURE (PUblic REquirements dataset), a dataset of 79 publicly available natural language requirements documents collected from the Web. The dataset includes 34,268 sentences and can be used for natural language processing tasks that are typical in requirements engineering, such as model synthesis, abstraction identification and document structure assessment. It can be further annotated to work as a benchmark for other tasks, such as ambiguity detection, requirements categorisation and identification of equivalent re-quirements. In the paper, we present the dataset and we compare its language with generic English texts, showing the peculiarities of the requirements jargon, made of a restricted vocabulary of domain-specific acronyms and words, and long sentences. We also present the common XML format to which we have manually ported a subset of the documents, with the goal of facilitating replication of NLP experiments.

This paper presents an approach for pragmatic ambiguity detection in natural language requirements. Pragmatic ambiguities depend on the context of a requirement, which includes the background knowledge of the reader: different backgrounds can lead to different interpretations. The presented approach employs a graph-based modelling of the background knowledge of different readers, and uses a shortest-path search algorithm to model the pragmatic interpretation of a requirement. The comparison of different pragmatic interpretations is used to decide if a requirement is ambiguous or not. The paper also provides a case study on real-world requirements, where we have assessed the effectiveness of the approach.