Are you still programming unit tests or already generating?

Your 80% code coverage goal is flawed. Learn how to generate tests based on black-box principles that truly improve software quality.

#1about 1 minute

Moving from manual unit testing to automated generation

The goal of test generation is to move from manual coding to a low-code platform that improves requirements coverage and collaboration between developers, testers, and requirements engineers.

#2about 2 minutes

Extending black box testing into the unit test level

The testing pyramid shows how black box testing, which focuses on requirements and interfaces, can be extended into the unit testing level, traditionally dominated by white box source code analysis.

#3about 4 minutes

Using equivalence classes and boundary values for testing

Black box testing relies on equivalence class partitioning and boundary value analysis to systematically test value ranges and their edges, where errors frequently occur.

#4about 2 minutes

A maturity model for unit testing beyond code coverage

A proposed maturity model for unit testing progresses from having no tests to implementing a full strategy with risk analysis and black box methods, moving beyond simple code coverage metrics.

#5about 2 minutes

Why 100% code coverage does not guarantee quality

High code coverage is the minimum requirement, not the goal, as it doesn't ensure that requirements are met or that the right tests have been written.

#6about 2 minutes

The workflow of the DevMate test code generator

DevMate parses a project's source code, stores its structure in an intermediate JSON model, allows users to define test data in a UI, and then generates standard, versionable test code.

#7about 7 minutes

Using systematic test data to find hidden code defects

A demonstration shows how to define valid and invalid test data based on requirements, automatically combine them into test cases, and generate code that reveals missing input validations.

#8about 2 minutes

Creating unit and integration tests with a visual canvas

Visit Test is a new standalone, web-based tool that provides a visual canvas for designing test scenarios, decision flows, and integration tests before generating code.

#9about 8 minutes

Q&A on test goals, AI generation, and tool specifics

The discussion covers setting better testing goals, the limitations of using ChatGPT for systematic test generation, how the tool handles code changes, and its local-only processing architecture.