Model-Based Testing Community | Connecting the MBT world…



Your Opinion about MBT

Dear all,

Model-based testing is good for …. wait. We had this several times before. What is model-based testing about? Isn’t it just the combination of several techniques that have proven their efficiency already?

In my opinion, we know already that modeling brings advantages because abstraction brings understanding, reusability, and a potential for automation. We also know that we have to test – at least if you care about the quality of your products (it’s even worse in the safety-critical domains) – and that every approach that reduces test effort is welcome.
So we know the advantages of models and of test automation. But are we also conviced of the powers of MBT?

At least sometimes I have the impression that MBT is not widely accepted, yet. Why? What are the reasons? Is MBT just located in an unfortunate part of the Gartner Hype Cycle? Are controllers never willing to spend money because they are paid to believe that we can make just this one project without MBT (…which means “without additional investments”)? Or is it just about trust? Do we need more success stories or more failure stories? I think that we had already several success stories … do we? Or is MBT accepted already and I just missed it? Prove me wrong. …or right. Or just tell me your opinion about it. 😉


No tags

1 comment

  • Geoff · June 5, 2014 at 7:37 pm

    If there are many success stories, the tellers of those stories have no awareness of this site. Or are they too busy with their heads down delivering the next product release. Truly unfortunate in either case.

    Marc-Florian’s presentation from the 2011 MBT UC laments the slow rate of adoption, despite the fact that there have been successes applying the UTP to MBT. My sense is that not much cultural change has occurred in the test domain. Look to the domain’s lexicon. Testers, Test Manager, Test Plan, Test Procedure, …

    Where are the rest of the requisite engineering roles and artifacts typically required to bridge from a system’s problem space to its solution space.

    A number of the tools have been slow to support and some possess support which is NOT methodology agnostic. So their implementation of the UTP as a language is less than fully conformant and interoperable with other tool environments.

    But the issue must be much deeper. My sense is that there is a fundamental difference in thinking about the design of a test system and a system’s test architecture than the thinking of the system’s design in the first place. Whether the system is a composition of hardware and software or software only.

    Perhaps, the lack of test requirements, as described by A. Wayne Wymore in his seminal work Model-Based Systems Engineering, is an indication of the root cause. Is Test Engineering considered as an integral element of systems engineering in system development universally?

    Despite the forward strides I see in efforts like that exemplified by the ISO 29119 series of standards as well as the OMG’s UTP efforts and ETSI TTCN3, there is considerable road ahead to changing culture.



Leave a Reply



Theme Design by
© by MBT Community 2011