Re: Conformance and Implementations

On Tue, Oct 09, 2001, Alex Rousskov wrote:
> On Tue, 9 Oct 2001, Rob Lanphier wrote:
 

> Thus, I suggest that W3C (QA) concentrates on testing (a practically
> useful aspect of protocol compliance) rather than certification.

I second that opinion. Starting a w3c certification process would be,
IMHO, costly (in terms of time and money), tricky, and in the end I'm
not quite sure it would be very efficient to addres our concern of
"bogus" compliance claims.

As for legal pressure on bad implementors who claim they're compliant, 
forget it. W3C has better to do with its time and money.

Let's test. Let's make good test suites, good implementation tests
and test reports. If we do that in such a way that people (users,
developers) will want to have a look at our reports when doubful about
conformance claims, and, later (?) when they want an overview of (good)
implementations, then we'll likely have the benefits of a certification
process (trust) without the hassle.

The main difficulty would then be to both stay informal and neutral
(opposed to official and/or flattering), and to be a valuable, accurate
and extensive source of information.


my 2 whatever.

-- 
Olivier

Received on Wednesday, 10 October 2001 02:08:45 UTC