[news.eclipse.tools.hyades] Re: Testability metrics

[mgn@xxxxxxxxxxxxx (Mike Norman)]

Magiel,

The hyades project should facilitate the sort of approach you are talking
about by defining standard ways that systems present their interfaces for
testing and standard ways of defining executable tests against those
interfaces, so that you can then have some abstract coverage metric which
is practicably computeable for a range of systems under test in the
context of a test suite or a generator of a test suite.   This said, I
think we're a little way from having the plumbing in place for you to
answer these kinds of questions in practical situations yet. The
definition of a test model (in UML) will help, and should be available in
a month or two.

Hope this makes sense,

Mike Norman,
Hyades project lead

Magiel Bruntink wrote:

> Hi,

> Currently I am researching into the topic of testability in the context of 
> my master's thesis. In short, a program's testability is the amount of 
> effort needed to satisfy a specified test coverage requirement during 
> testing. Much of the literature focuses on defining measures of the source 
> code (for example: failure hiding ability, domain-to-range ratio, coupling, 
> etc.) and using these to draw conclusions about a program's testability.

> The collection and presentation of these measures is something I'de expect 
> to be useful in the Hyades framework. For example, knowledge of a program's 
> testability can play an important role in evaluating the progress of the 
> testing effort.

> Are there existing plans to implement any source-based measurement tooling 
> in Hyades? Would it be possible to contribute to such a project?

> Yours,

> Magiel Bruntink,
> University of Amsterdam (mbruntnk@xxxxxxxxxxxxxx)
> Software Improvement Group (magiel@xxxxxxxxxxxxxxxxxxxxxx)