End-to-End Rationale Reconstruction

TL;DR

This paper proposes Kantara, an end-to-end ML and NLP-based pipeline that automatically extracts and constructs a knowledge graph of design decisions and rationales, ensuring traceability and correctness, demonstrated on Linux Kernel data.

Contribution

It introduces Kantara, a novel context-independent approach for automatic rationale extraction and knowledge graph construction in software design.

Findings

Preliminary evaluation shows promising results on Linux Kernel data.

Kantara effectively models decision evolution and traceability.

Validation mechanisms improve information correctness and coherence.

Abstract

The logic behind design decisions, called design rationale, is very valuable. In the past, researchers have tried to automatically extract and exploit this information, but prior techniques are only applicable to specific contexts and there is insufficient progress on an end-to-end rationale information extraction pipeline. Here we outline a path towards such a pipeline that leverages several Machine Learning (ML) and Natural Language Processing (NLP) techniques. Our proposed context-independent approach, called Kantara, produces a knowledge graph representation of decisions and of their rationales, which considers their historical evolution and traceability. We also propose validation mechanisms to ensure the correctness of the extracted information and the coherence of the development process. We conducted a preliminary evaluation of our proposed approach on a small example sourced from the Linux Kernel, which shows promising results.