[SystemSafety] MC/DC coverage assumptions

[Derek M Jones]

Dewi,

>> This sounds like bad user interface design.  Users could be
>> led to believe the numbers are for MC/DC, when in fact they need
>> to merge in other values to get this data.
>>
> 
> That's only true if you're trying to measure percentage coverage. When
> following DO-178C, we're only interested in whether we can justify any gaps
> in coverage. So in your example where a decision with a single condition

I appreciate that the presentation of information makes sense in
a common use context, but in other contexts it can be misinterpreted.

Misinterpretation of information presentation is a cause of accidents.
There was one such discussed recently on this list.

> had only taken one outcome, the gap in decision coverage would have to be
> resolved as explained by Tom Ferrell. There's no compelling reason for the
> tool to point out the same gap in MC/DC, because the resolution would be
> the same.
> 

-- 
Derek M. Jones           Software analysis
tel: +44 (0)1252 520667  blog:shape-of-code.coding-guidelines.com