Software Requirements Data Set

For automatic software requirements specification (SRS) processing and quality evaluation, we have collected textual requirements from four different sources:

a set of requirements from the PUblic REquirements (PURE) dataset (folders prefixed with upper case 'P') [https://zenodo.org/record/7118517],

a set of requirements collected by Hayes et al. (folders prefixed with upper case 'H') [https://doi.org/10.1109/REW.2019.00052],

a set of requirements provided by Dalpiaz (folders prefixed with lower case 'g') [https://doi.org/10.1016/j.infsof.2018.12.007],

a custom collection of requirements found in the Internet (folders prefixed with upper case 'C').

We preserve the original file of the requirements (TXT, PDF, RTF, DOC or HTML) and add a plain text conversion (files called _Raw.txt) or, for some of them, Markdown form (files called _Raw.md). For some of them, we have extracted part of functional requirements (files called FunctionalRequirements) or functional specification relevant parts (files called AllRelevant).