W3C home > Mailing lists > Public > www-qa@w3.org > April 2002

RE: Outsourcing in QA and Testing

From: David Marston/Cambridge/IBM <david_marston@us.ibm.com>
Date: Mon, 15 Apr 2002 22:04:35 -0400
To: www-qa@w3.org
Message-ID: <OF57A7EC5A.90645CA9-ON85256B9D.0009F1A6@lotus.com>

Brian Kelly wrote:
>How about testing of Web sites against browsers rather than against
>standards (W3C are interested in the latter, but users in the former).

How about testing of browsers against the standards, then? That might
be a lost cause (at least until XHTML is widespread), but it's the
kind of issue that is suitable for this list. The current idea goes
approximately like this:
1. A Working Group (WG) wants to specify a vocabulary, protocol,
process, API, or some other thing.
2. The QAWG wants WGs to identify "classes of products" that will
act upon the thing from (1) above.
3. The QAWG wants WGs to have sufficiently rigorous conformance
verbiage in the specs so that an "independent" test lab could assess
how well a product conforms to the specs. For this context, I define
"independent" as: having no WG members or other informal ties to the
WG that wrote the spec, so that they are relying solely upon the
materials published by the W3C (and other specs cited as normative).

I think that the Web sites themselves don't fit directly into the
above model. They are content or "instance documents" and the QAWG is
more concerned with the browsers and other tools that produce or
consume the HTML. A conformance test suite would contain synthetic
instances of documents that exhibit all the correct and incorrect
content.
.................David Marston
Received on Monday, 15 April 2002 22:07:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 6 December 2009 12:13:59 GMT