W3C home > Mailing lists > Public > www-dom-ts@w3.org > April 2001

SV: [Xmlconf-developer] FW: [General] Language-independent test representation

From: Dimitris Dimitriadis <dimitris.dimitriadis@improve.se>
Date: Mon, 30 Apr 2001 10:49:22 +0200
Message-ID: <9F67DC27F4CCD311ABA600508B6A66A44A61EB@VXOIMP1>
To: "'David Brownell'" <david-b@pacbell.net>, xmlconf-developer@lists.sourceforge.net, www-dom-ts@w3.org
Hi Dave

Thanks for you comments.

Currently, the vocabulary is not that rich, ie. it covers basically the test
itself. We are thinking on terms of prvoding information about expected
results and so forth, especially to be able to generate advanced results
(currently the output is a colour-coded table) but have already started
thinking of a more advanced set of alternatives.

The DTD for this vocabulary will be posted to the list shortly in order to
be discussed.

Regards,

Dimitris

-----Ursprungligt meddelande-----
Från: David Brownell [mailto:david-b@pacbell.net]
Skickat: den 28 april 2001 20:09
Till: Arnold, Curt; xmlconf-developer@lists.sourceforge.net
Kopia: www-dom-ts@w3.org
Ämne: Re: [Xmlconf-developer] FW: [General] Language-independent test
representation


> 1. The tests will be represented in a language-neutral, XML format.
> This format will be posted to this list to be discussed 

I'd at least like to see some standards for the test metadata.
That is, XML would naturally be used to describe each test case,
its input data, its output data, and how results get reported.

I have in mind that a test run could just spit out lots of XML
results (test ABC passed, failed with result "foo", etc.) that an
XSLT stylesheet would crunch into a test results report (which
could highlight nonconformance problems).  In the same way,
some different stylesheet could crunch a test description report
from those test inputs, which could be used to assess testing
depth and coverage.

That perspective leaves open a hole in the middle where
the language-specific API testing is actually done, which I
suspect is where some of the current thought has gone.  For
that I'd likely be content to have collections of language
specific code making the API calls -- which may not be
the model that some other folk are working with.

- Dave
Received on Monday, 30 April 2001 04:54:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:34:02 UTC