RE: testability definition

Yep it is exactly the difference. But I don't think we're that far from
each other.
I see the goal of the SpecGL is to facilitate that the W3C specs are
unambiguous and define the behavior for every input.

IMO "practical testability" is:
- Definition of the testability (theory) 
- guide how to improve specs towards testability (practice)

Tx

-----Original Message-----
From: Alex Rousskov [mailto:rousskov@measurement-factory.com] 
Sent: Friday, September 06, 2002 4:17 PM
To: Kirill Gavrylyuk
Cc: www-qa@w3.org
Subject: RE: testability definition

On Fri, 6 Sep 2002, Kirill Gavrylyuk wrote:

> "A requirement is testable if a finite cost-effective process
> 	 can detect any known a priory violation of the requirement.
> 	 A specification is testable if all its requirements are
> 	 testable."
>
> The problem is that it is not clear how do you deduce the
> requirements from a spec.

This is a completely different issue and is irrelevant here. Yes,
there are specs without clear requirements (e.g., HTTP/0.9 and
HTTP/1.0). There are specs without the definition of valid input! All
those are not testable in the context of SpecGL. Some of those should
not be even called "spec".

> You need to be able to build a process that would deduce all the
> requirements from the spec.

It is not a spec if you cannot deduce the requirements. Or, at least,
it is not a spec worth talking about. It does not matter whether it
will be SpecGL-testable or not.

>  There are buggy untestable specs that
> would satisfy the definition above just because they do not specify
> the behavior for some inputs.

Those specs have other problems. I do not see any harm to say that a
spec that does not define behavior for some inputs is testable. That
spec will violate other SpecGL requirements. The testability
definition should only be concerned with requirements that can be
identified, IMO.

> Consider a testable behavioral spec as a many:1 map between an input
> space and a behaviors space.
>
> The spec is in trouble if either >1 behaviors are specified for the
same
> input or none of the behaviors are specified for an input.
>
> By enumerating the input space you can prove the testability.
>
> Thoughts?

I am talking about practical testability. You are talking about
theoretical testability and assume that all input spaces can be
enumerated. If you ignore the latter bug, both approaches are valid. I
just do not think that SpecGL should care about theoretical
testability if it is not also practical.

Alex.

-- 
                            | HTTP performance - Web Polygraph benchmark
www.measurement-factory.com | HTTP compliance+ - Co-Advisor test suite
                            | all of the above - PolyBox appliance

Received on Friday, 6 September 2002 20:10:59 UTC