Improving requirements with NLP techniques

Abstract

Elaborating “good” requirements specifications is an important factor for the success of a software project. Requirements are normally expressed using textual descriptions in natural language, but not without problems. Some requirements documentation techniques, such as use cases specifications, often focus on functionality and leave many concerns understated in the text and scattered through several documents. These concerns, commonly known as crosscutting or architecturally-relevant concerns, often come from business goals or quality attributes that must be clearly identified by analysts and developers, as these concerns can have a far-reaching effect in the development process. Not treating these concerns at early development stages can lead to poor design solutions that become difficult (and costly) to fix afterwards. Unfortunately, searching for concerns in textual requirements is a difficult and time-consuming task for analysts, because requirements are often poorly modularized and there is text duplicated across documents.

<i>(Párrafo extraído del texto a modo de resumen)</i>