Techniques based on the software engineer’s experience and intuitions are as follows:
Ad hoc testing
Probably the most widely used technique. These tests are based on the experience, intelligence, intuition and knowledge of an engineer who considers the problem from the standpoint of previous analogies. This type of testing can be suitable for defining those tests that are not covered by the more formalized techniques.
Exploratory testing
This testing involves doing 3 types of activities at the same time: training, test design and test execution. This type of testing is not preliminarily defined or included in the test plan and such tests are written, executed and corrected on the fly, as necessary. The effectiveness of exploratory tests directly depends on the knowledge of the engineer, formed based on the behavior of the product being tested, the degree of familiarity with the app, the platform, the types of possible defects and failures, the risks associated with a particular software product, etc.
Be advised that exploratory testing services are provided to learn about the product while executing tests on it. In so doing, it is possible to find out what the app does and what it does not do, learn its features and make sure that it works for its users.
Specification-based techniques
Equivalence partitioning
The app’s area under consideration is divided into equivalent classes that are considered equivalent in terms of the relationships on hand and characteristics of the . A representative set of tests (sometimes just one test) is created of equivalence class tests (or sets of classes).
Boundary-value analysis
Tests are built with the focus on the use of those quantities that determine the ultimate characteristics of the system under test. Continuation of this technique is robustness testing of the system, conducted with values that go beyond the specified limits of values.
Decision table
Such tables represent logical relationships between conditions (can be considered as “causes”) and actions (can be considered as “effects”). A set of tests is built by sequential consideration of all possible cross-relationships in such a table.
Finite-state machine-based tests
They are developed as a combination of tests for all states and transitions between states represented in the corresponding model (transitions and states of the application).
Comments are closed.
