Key Points

• Most test programs are focussed on verifying that software conforms to the specification, and operates without error under expected conditions.
• However, functional combination means that there is almost an infinite number of possible tests for any given application, and that the vast majority of these tests will be of little value.
• Additionally, the assumption that testing based on functionality will be sufficient is incorrect û in many software projects a full functional test program will often cover less than 60% of the code.
• Other techniques must be utilised to improve the validity and completeness of functional tests, without adding a huge burden of new tests.

Presentation Abstract

Functional testing is often thought to be a straightforward logical process of converting functional specifications into test scenarios.

However there are several major challenges faced by functional testers.

The initial challenges are associated with the task of deriving a full functional test set and verifying that functional testing is complete. These challenges are in fact impossible to overcome because functional combination means that there are almost infinite scenarios which can applied during testing. It is therefore crucial to ensure that a minimum set of functional tests is executed providing maximum benefit.

Secondly, functional testing is very frequently under considerable time pressure due to slippage in earlier stages of the development process. This means that testing often has to be cut short in order to deliver software on time. It is therefore critical to ensure that tests are prioritised with important tests being performed early.

Additionally, the architecture and implementation of a software project influences the nature and scale of the required testing.

How can we improve our testing to ensure that the critical parts have been adequately tested without doubling or trebling the number of tests required?

McCabe & Associates promote powerful software analysis techniques and tools to assess the risk associated with a particular test schedule and identify necessary improvements. These techniques will be outlined during the presentation.

About the Speaker

Matthew Brady has worked for McCabe & Associates in the UK distribution office for four years. During that time he has been involved in the implementation of McCabe solutions in many organisations across Europe including the UK National Air Traffic Centre, NEC, Nortel, Telefonica and Pace. He now heads up the consultancy and support team for McCabe in Europe.

Prior to McCabe, he worked for Kinesix Corp. delivering real-time graphics solutions to the aerospace and safety critical markets.

He graduated in 1987 in Mathematics from Manchester University