Early versions in software development

Abstract

The development of early versions in software plays a crucial role in the software engineering process, allowing teams to test concepts, validate ideas and collect user feedback before committing to full-scale development. However, the terminology used to describe early versions, such as pro- totype, Minimum Viable Product (MVP), Proof of Concept (PoC) and increment, varies and the terms lack a clear scientific definition. Therefore, the purpose of this research is to see what distinguishes early versions from each other, and if and where they overlap, by making a scientifically grounded framework.

To achieve this, a literature review was conducted, analyzing the definitions of early versions from academic sources and grey literature. Additionally, perceptions of individuals knowledgeable in the field of software engineering were gathered through interviews and a classification exercise to understand practical distinctions between these terms.

The findings reveal that existing definitions are sometimes overlapping or context-dependent. However, by identifying purpose, maturity and other characteristics of early versions, this study proposes two frameworks to differentiate early versions: (1) a TRL-based timeline and (2) a purpose-based classification. The timeline aligns early versions with Technology Learning Levels (TRLs), mapping PoCs to TRL 2-3, non-working prototypes to TRL 4, working prototypes to TRL 5, MVPs to TRL 6, alpha versions to TRL 7, beta versions to TRL 8, and final 1.0 versions to TRL 9.

This classification highlights how different early versions serve specific goals at various stages of readiness. The purpose-based classification groups early versions into three categories: experimental (PoCs), exploratory (prototypes and MVPs), and evolutionary (increments, MVPs, and prototypes).