This is the supplementary material repository of the paper "A Taxonomy for Design Decisions in Software Architecture Documentation". In this paper, we propose a taxonomy for design decisions in software architecture documentation with the purpose to support activities, such as specific test case creation, trace link recovery, or the identification of inconsistencies between different artifacts. Additionally, we developed a new dataset comprising 17 software architecture documentations from publicly available open-source projects. We classified the contained design decisions based on the taxonomy. Further, we explored if state-of-the-art machine learning-based classifier are able to automatically identify the design decisions and their subcategories in the dataset.

This repository contains the dataset and code used in the paper, as well as information on the annotation process:

• dataset contains the extracted software architecture documentations (.txt files) and the per-line labels (.csv files)
• annotation contains information about our labeling:
  • texts: the extracted software architecture documentations
  • annotator1: the results of the labeling of the main annotator for all 17 documentations
  • annotator2: the results of the first annotator involved in the labeling of the three projects used for manual validation
  • annotator3: the results of the second annotator involved in the labeling of the three projects used for manual validation
  • IAAgreement: inter-annotator agreement (Krippendorf's α, KALPHA)
• implementation contains the python modules and notebooks used to produce the results of the automated classification
• results contains the results of the automated classification

The following tables shows examples for the leaf classes.