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

RE: DOM TS copyright and sf at w3.org

From: Arnold, Curt <Curt.Arnold@hyprotech.com>
Date: Fri, 8 Jun 2001 12:25:00 -0600
Message-ID: <B2C1451A181BD411B88A00E018C1C19C08ACE6@thor.aeathtl.com>
To: "'www-dom-ts@w3.org'" <www-dom-ts@w3.org>
>> Should each test
>> have a "Copyright (c) 2001, MIT..." boilerplate?
> 
> Yes, they should.

I thought so.  We should provide a standard formulation.
Hopefully, it would be sufficient to place something like:

<!-- Copyright (c) 2001, whoever.  This document is made
available under the W3C Document Notice http://www.w3.org/...  -->

instead of having to include the full text of the document
notice in every test.

A transform could be provided that inserts the appropriate
boilerplate when the proper formulation is decided (but 
before the submission process is started)

> No. The tests will be endorsed by the W3C and therefore, you 
> won't be allowed
> to modify them and still claim conformance with the test 
> suites. NIST and W3C
> agreed that the W3C Document Notice and License is the 
> appropriate one. Note
> that I said the tests, not the framework used to produce them. The
> user/developer must be able to download the tests from the 
> W3C web site
> http://www.w3.org/DOM/Test and check his implementation for 
> DOM conformance.

I understand the need to keep the official tests clean, but 
I see an conflict between the expected use of the tests
and the "No right to create modifications or derivatives 
of W3C documents is granted pursuant to this license."

I could see the following scenarios that seem like a legitimate
use of the tests but progressively get closer to conflicting
with the prohibition on derivative works.

1. Test interpreter

A third party provides a test interpreter that executes the tests 
by reading the official W3C test archive
and invoking methods in the implementation through Java reflection,
IDispatch or similar mechanism.  This seems like the cleanest
use of the tests, but most difficult to implement, especially
on platforms where there is no built-in support for reflection.

2. Test compiler

A third party provides a process that converts the official
W3C test archive into equivalent source code for the combination
of language, test framework and processor being used.

Running the test would require the test definitions,
some coordinator (like Ant or Make),
a compiler (like javac), in addition to the test framework.

The provider of this tool would not be infringing on the derivative
work clause but would be assisting the person running the tool 
to infringe since the process would create Java, C++ or 
other source code that is derived from the W3C document.

3. Compiled tests

A third party processes the official W3C test archive 
to create equivalent source code, compiles and 
produces a component (.jar, executable, 
web site) that will test a particular processor.

Running the test would not require the test definitions,
a coordinator or a compiler.

In this scenario, there would appear to be a 
potential conflict with the derivatives clause 
since the component contains compiled code that 
was generated from the tests.

In all three scenarios, the provider of the interpreter
or translator could corrupt the results of the tests
by partially or incorrectly interpreting the test
definitions.

4. A test author wants to use a W3C DOM test as the
basis for a new test

For example, I have quite a few tests for DOM Events
that are deriviatives of the NIST DOM 1 tests.  Basically,
I've taken tests in the NIST DOM 1 suite that 
made mutations to the DOM, added event listeners and checked that
the events that I personally expected occurred.

The tests are not trying to masquerade as the original test
or imply any endorsement from NIST, but they are clearly
derivative works.  The Software Notice would allow this type of
use (and prohibit passing off the derivative as the original)
but this would seem to be a clear violation of the Document
Notice.

5. A framework author wants to use a transform that produces
Java code from the test definitions as the base for a transform
that produces code for some other language.

If the transform was under the Document notice, this would seem 
to be a violation but would be acceptible under the Software notice.

> There is a Bugzilla 
> system in the W3C
> but we're still testing it so it is not ready yet.

If you are already working on Bugzilla, then it would probably
be best to keep the resources focused and get it usable
for this project instead of bring another piece of software
into the mix.
Received on Friday, 8 June 2001 14:36:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 6 April 2009 12:58:44 GMT