Thoughts on Testing

As of May, 2005, NodeBrain has still not been tested in a rigorous systematic way. It was built as a prototype, found to be useful, and then successfully applied to a couple specific problems. For the purposes of our own applications, we have gotten by with random parallel tests as new versions have been introduced. This was normally, but not always, preceded by simple contrived tests that focused only on the intended changes.

As an open source tool, nb clearly requires more complete testing before each release. The links below outline some common general approaches we hope to implement as part of the release process. This is just intended to give testers ideas to expand on in building actual tests.

• Random Parallel Tests
• Controlled Automated Tests
• Sample Tests