RE: Ideas collected during the W3C QA Workshop

The number of attributes returned, for example using a HTML document and
a <DIV> tag depends on the respective DTD. I completely agree with you
that the test case needs to specify, which one to use if a certain
number is expected. As a minbar I would be satisfied if the attribute
collection exists and returns workable values.

-- Markus

-----Original Message-----
From: Joseph Kesselman [mailto:keshlam@us.ibm.com] 
Sent: Monday, April 09, 2001 12:37 PM
To: Markus Mielke
Cc: www-dom-ts@w3.org
Subject: Re: Ideas collected during the W3C QA Workshop


>I think the issue being addressed here is that tests shouldn't expect
>elements in any particular order

Watch out for terminology, since "element" is a key word when discussing
XML... A test should certainly expect Elements to be in a particular
order,
since that's semantically meaningful and well-specified. There are
specific
kinds of Nodes that are not considered to be ordered, Attrs being the
most
obvious, and a test for those specific cases should confirm that the
right
things are appearing but not enforce order.

>web based issue (bug) tracking system

Apache's started using Bugzilla again, now that some security and
configuration concerns have been dealt with. I have no opinion on it
versus
other tools, but so far it seems to work and some folks may already be
familiar with it.

>The number of returned attributes is dependent on the DTD of the given
>platform.

I don't think I understand this issue... The DTD's default attributes do
get included in the result, certainly. But the DTD should be specified
as
part of the testcase, so that shouldn't be an issue. If you may be
running
against multiple DTDs (eg for HTML), either specify which one much be
used
for this test or enumerate what the results should be for each.

______________________________________
Joe Kesselman  / IBM Research

Received on Monday, 9 April 2001 16:20:57 UTC