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

Re: [General] DOM Level 1 HTML tests

From: Mary Brady <mbrady@nist.gov>
Date: Fri, 21 Dec 2001 10:13:02 -0500
Message-ID: <011601c18a31$fbcf4520$293b0681@sdct.nist.gov>
To: "Arnold, Curt" <Curt.Arnold@hyprotech.com>, "Dimitris Dimitriadis" <dimitris@ontologicon.com>
Cc: <www-dom-ts@w3.org>
comments inlined

--Mary

----- Original Message -----
From: "Dimitris Dimitriadis" <dimitris@ontologicon.com>
To: "Arnold, Curt" <Curt.Arnold@hyprotech.com>
Cc: <www-dom-ts@w3.org>
Sent: Friday, December 21, 2001 3:50 AM
Subject: Re: [General] DOM Level 1 HTML tests


> comments inlined
>
>
> On Thursday, December 20, 2001, at 10:57 PM, Arnold, Curt wrote:
>
> > Philippe wrote:
> >> To be precise: the current DOM Level 1 Core test suite only
> >> tests the DOM XML implementation in Mozilla and IE. It
> >> doesn't test their DOM HTML implementation. Curt tested the
> >> DOM SVG implementation of Adobe/Batik against the DOM Level 1
> >> Core Test Suite. We should be able to do the same to test the
> >> DOM HTML implementation. After all, a DOM HTML implementation
> >> must support the Core at the same time.
> >
> >
> > It was fairly simple to take staff.xml and create staff.svg since SVG
> > allows
> > any non-SVG namespace element to appear in SVG and is ignored by the
> > reader.
> > The only change that was significant to the tests was that the root
> > element
> > was "svg" instead of "staff" causing a handful of tests to fail.
> >
> > The DOM L1 tests, though not explicitly tied to a contentType, do
> > require
> > that there are elements named "employee", "address", etc that are not
> > representable in traditional HTML.
> >
> > I had been waiting for NIST to commit their HTML tests for several weeks
> > now.  My suspicion is that some significant portion of these may really
> > be
> > Core tests working on HTML content instead of tests that depend on HTML
> > interfaces.
> >
> [dd] Do you mean the Level 1 or Level 2 tests?

[mb] I'm not sure that I agree with your suspicion.  The HTML tests are for
the HTML
module itself.  It is true that in order to get to the module specific
interfaces, you
need to use some of the core interfaces.  I would not necessarily presuppose
that
all of the core interfaces are fully exercised.

>
> > As I suggested in a previous message, it would be pretty simple to
> > determine
> > those tests that fall into that category by attempting to validate those
> > files against the DOM L1 schema (with the namespace hacked to
> > Level-2).  If
> > a test validates against the hacked L1 schema, then it does not depend
> > on
> > HTML specific interfaces and should be moved to Core.
> >
> > I'd would recommend holding off finalizing DOM L1 core until the NIST
> > HTML
> > test submission is reviewed.
> >
> [dd] I'd like to wait and see what chances we have of getting Level 1
> HTML tests before holding back the release. In any case, there's been
> plenty of time to sumbit tests, and since this is a publically developed
> framework, perhaps there is not enough interest in Level 1 HTML for
> tests to be submitted.
>
[mb]  The NIST HTML submission is for Level 2 HTML -- HTML Module
only.  As it turns out, many of these tests could also be for Level 1 HTML
as
well.  I expect that there probably is interest in having Core tests for
HTML-only
implementations -- the biggest problem is the resources to actually write
the
tests.  We will submit some the the HTML Module tests later today, but will
not
be able to look at Core HTML until after the holidays.

> > I'm starting to travel tomorrow and will be infrequently checking my
> > email.
> > Hope that all of you have a great Christmas and New Years.
> >
>
> [dd] Thanks, same to you.
>
>
Received on Friday, 21 December 2001 10:13:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 6 April 2009 12:58:45 GMT