LEMON: a Wizard to Leverage the Use of Requirements Patterns in Automatic Software Generation Tools - BNF, survey demographics, raw data, and threats to validity
David Mosquera, Marcela Ruiz, Michael Wahler, Christen, Matthias, Jürgen Spielberger, Óscar Pastor · Zenodo (CERN European Organization for Nuclear Research) · 2023
This repository contains the LEMON BNF, survey demographics, raw data, and threats to validity, among other resources from the paper "LEMON: a Wizard to Leverage the Use of Requirements Patterns in Automatic Software Generation Tools", currently under review at the industry innovation track at RE'23. How to use: In this repository, you will find three files. We describe each file in the following numerals: ConcreteSyntax-LEMON-RE23.pdf. It contains LEMON requirements patterns specification language BNF (Backus-Naur Form) https://en.wikipedia.org/wiki/Backus–Naur_form. You can find all statements defined and an example (not exhaustive) for using the language. Pattern-based requirements engineering survey.xlsx. It contains raw data from online survey results. You can find the following sheets inside this MS Excel file: Demographics. It contains anonymized demographics from 11 subjects that participated in our survey. You can identify each subject by an ID followed by a column for each demographic question. Survey-raw-data. It contains the answers from 11 subjects to our survey questions about LEMON's potential adoption in practice. You can identify each subject by an ID followed by a column for each perception question. Traceability. It contains the traceability between survey-raw-data and the challenges we describe in the paper. You can find in the first row the names of each challenge followed by an answer from our subjects clustered in such a challenge. We performed such clustering after analyzing the raw data. SurveyQuestions. It contains a preview of our survey. You can find a set of screenshots from Google forms. Moreover, you can find a link to check the survey yourself. Threats to validity: In this section, we discuss potential validity threats in our survey. Internal validity threats. Instrumentation validity: In order to ensure internal validity, we carefully considered the instrumentation used in our study. We chose an online survey out of several empirical research instruments available. In future studies, we plan to collect data through multiple channels, including directly from industrial organizations, to avoid any potential threats to internal validity related to relying on a single channel. Sampling validity: We acknowledge that the sample used in our study was not truly random, as it was limited to the author's LinkedIn connections and SHIFT consortium participants. To address this potential threat to sampling validity in future studies, we plan to distribute our survey through general venues where requirements engineers, healthcare professionals, and software engineers can participate. External and construct validity threats. Results generalization: We recognize that our sample size was small and may not be representative of the population as a whole. To improve the results generalization, we plan to replicate our study with larger sample sizes and more diverse populations. Results accuracy: We acknowledge that our study included participants with varying levels of experience, ranging from 0 to more than 5 years of experience in requirements engineering. To further evaluate how opinions may vary based on experience, we plan to conduct future studies with larger sample sizes that account for this potential threat to construct validity. If you have any doubts, contact [email protected] / [email protected] / [email protected]