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

Re: proposed Test Materials license

From: Lofton Henderson <lofton@rockynet.com>
Date: Fri, 15 Nov 2002 15:15:06 -0700
Message-Id: <5.1.0.14.2.20021115092224.025eb220@rockynet.com>
To: reagle@w3.org, "Kirill Gavrylyuk" <kirillg@microsoft.com>
Cc: <www-qa-wg@w3.org>, "Karin Rivard" <rivard@MIT.EDU>, "Marija V. Jankovich" <marija@MIT.EDU>, Philippe Le Hegaret <plh@w3.org>

At 04:40 PM 11/14/02 -0500, Joseph Reagle wrote:

>On Thursday 14 November 2002 04:06 pm, Lofton Henderson wrote:
> > One problem with the Software License is that it places no constraints on
> > the types or extent of modifications that may be done.  Users of a
> > W3C-distributed Test Suite absolutely should not be allowed to make
> > modifications to the substance of the tests.  This is why the
> > "Operational Guidelines" recommend the Document License.  However, as
> > people have noted, modifications to the framework, harnesses, operating
> > interfaces, etc, may be necessary for some kinds of test suites.
>
>If someone did want to modify the tests they would have to document the
>changes.

Ref:  http://www.w3.org/Consortium/Legal/copyright-software-19980720

For my taste, the restrictions are too weak (Ref. #1 and #3), and 
potentially too easy to circumvent the spirit of.  I.e., I could 
technically comply with #1 and #3, while my notices are sufficiently vague 
or obscure that only the fastidious will find and understand them and their 
implications.

For example, if the SVG test suite were under Software License (it is under 
Document), I could make a modified version of the SVG test suite that 
suppresses features that my implementation fails.  I believe that I would 
satisfy #3 by putting some obscure comment
(XML comment, <!-- ... -->) in the SVG source code.  The naive observer 
would see a test displayed -- apparently a member of "W3C SVG 1.0 Test 
Suite, Release 2.0" -- but would not know the truth about it unless he/she 
happened to "View Source" (and be technically competent).

Notice that I'm assuming deceptive intent -- it happens occasionally.

I'd be interested in some interpretations or some "for examples" -- what is 
acceptable and what is unacceptable for #1 and #3 (under a strict reading 
of those)?

>To further this, the trademark restriction was added to preclude
>any misrepresentation that the modified form was in any way sanctioned by
>the W3C.

(... are you referring to the last paragraph of the Software License?).

That's good.  So at least someone who wants to publicly claim, "we pass the 
FOOBAR Test Suite" (who is using a modified version) cannot legally do so 
without W3C scrutiny and permission.

>So for instance if we released the W3C FOOBAR Test Suite under the
>W3C Software License, others could:
>1. Modify the framework/harness to their context. If they republish they
>have to document the change of course and can not represent it as "the" W3C
>test suite any more.

This is our conventional "need to modify" scenario.  It leaves intact the 
test assertions (and pass/fail criteria) of the test, and allows 
environmental modification and adaptation.

>2. If someone was interested in tracking the errata, or versioning them to
>the next version (if for instance, this wasn't being done by a WG), they
>would be permitted to do so, but again they would have to document the fact
>and not incorrectly represent the test's authenticity and standing with
>respect to the W3C.

This is an interesting scenario that *would* require modifying the 
underlying test assertions associated with tests.


>I'd hope this would satisfy your concern regarding the standing of the test
>suites.

For my taste, for W3C-hosted test materials, I would like the option of 
stronger constraints than I'm reading into the Software License.  (Unless 
I'm misreading it.)

This is distinct from Kirill's other problem -- corporate policy that 
prevents donation of test materials without scope-of-use restrictions.

Regards,
-Lofton.
Received on Friday, 15 November 2002 17:14:43 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Thursday, 9 June 2005 12:13:11 GMT