W3C home > Mailing lists > Public > www-qa-wg@w3.org > February 2003

Re: DRAFT minutes, QA Working Group Teleconference 2003-02-24

From: Joseph Reagle <reagle@w3.org>
Date: Wed, 26 Feb 2003 10:09:10 -0500
To: "Kirill Gavrylyuk" <kirillg@microsoft.com>, <www-qa-wg@w3.org>, "Lofton Henderson" <lofton@rockynet.com>
Message-Id: <200302261009.10391.reagle@w3.org>

On Monday 24 February 2003 22:51, Kirill Gavrylyuk wrote:
> 1. Warranties and Liabilities.
> The W3C Software License provides no warranties against the materials
> (software) containing viral code, or infringing someone's patent or
> copyright rights, etc. When allowing its employees to use the software
> published under such license, a company may end up redistributing these
> viral or infringing materials. This will cause damage to its customers
> and therefore the company itself. This makes it hard to asses the risks
> of allowing the use of the materials published under such license. Why
> then to contribute to something that you cannot allow your employees to
> use?

I'm still having a difficult time understanding this one Kirill:
1. Not providing warranties is a standard licensing term. I think we agreed 
to that on the call and would not expect the W3C to do otherwise. (As an 
aside, when someone makes a contribution to a test suite we ask the 
contributor if state they believe they have all the rights necessary to 
make the contribution [1].)
2. If a company redistributes software with the test suite they too probably 
waive as much liability and warranty as permitted by law with their own 
associated software.
3. There is nothing *forcing* an organization to redistribute a test suite. 
We have reports that folks will want to and *do* distribute test suites as 
part of their software. Furthermore, if the redistribution is important to 
them and they *want* to make additional warranties, then all the more power 
to them.
4. An organization that wants its employees to use a test suite, can have 
its employees use the test suite. Again, nothing forces them to 
redistribute it simply because it was on an engineer's desktop.
5. If an organization develops a set of tests, which are contributed to the 
W3C suite, they still have the rights to their original tests to 
orthogonally adapt or redistribute them as they see fit.


[1] http://www.w3.org/Consortium/Legal/2002/contribution-grant-20021231

> 2. Scope of use.
> I think we went pretty light on this today, but I believe this is one of
> the key reasons why Microsoft would not like to make contributions to
> the materials published under the W3C Software License. We are happy to
> donate test cases for free to be used for the purpose they were created.
> Hence the desire to control the scope of use. Why would we donate them
> for free to be used for anything else but testing? Document License
> provides us a limited scope ("fair use") for the test cases which is
> close. That's why we were happy donating a large number of test cases
> for Schema and SOAP to be published under the Document License so far.

I can understand this and accept that some organizations may wish to limit 
the use of their contribution by others -- other than the W3C for that 
specific activity.  Though, note that the Document License does not govern 
"use" or limit its scope. It prevents derivation/adaptation. To use my 
previous example, I can print out the schema/soap test cases and use them 
to raised the height of my monitor. <smile/>

> 3. Redistribution. This was mentioned mostly to mitigate the risk
> expressed in 1. Therefore I am not that concerned about redistribution
> of the test cases/documentation, but only test software.

Understood, see above with respect to warranties.

> 4. Modification/derivation work. Patrick brought up interesting
> scenarios of redistributing/incorporating the W3C test cases into a
> proprietary bigger test suite. Thinking on it a bit during lunch, I
> would call a fair use of a W3C test case collection the following:

We have to be careful with "fair use." ("Use" in this context can be 
confusing since I've been saying that copyright doesn't govern usage 
itself, but in this context "use" should be read as an exercises of one of 
the enumerate exclusive rights, nothing else.) It provides limited 
exemptions to the exclusive rights for purposes such as, "criticism, 
comment, news reporting, teaching (including multiple copies for classroom 
use), scholarship, or research" with consideration of the factors of the 
purpose/character of the "use", nature of the copyrighted work, the amount 
use, and the effect on the market of the copyrighted work.
  http://www4.law.cornell.edu/uscode/17/107.html

But I agree with you and Patrick that the issues below do need to be 
considered. While I, hesitatingly, mentioned that making a local 
language/platform specific adaption might be considered fair use, the 
addition, subtraction, annotation of errors and such would probably be 
reaching on that point. (Best to have specify exceptions in the FAQ or use 
a Test License.)

> 4a. Addition. Add company's own test cases - clearly separate the W3C
> test suite and company's test cases. You can add a W3C test collection
> only as a whole, cannot exclude any of the test cases from it. (I wasn't
> sure about that during the telconf, but convinced now)
>
> 4b. Mark errors. Annotate test cases that they believe are in error in a
> separate document. Do not exclude those test cases. Do not edit
> them/change their context.
>
> 4c. Adding more metadata to existing test cases. If needed, add
> additionally a metadata for the test cases, by placing a separate copy
> into a separate folder (or building it dynamically on the fly). For
> example, in case of XML Schema test collection, if one needs to add more
> attributes then provided by the control file, one could supply a control
> file with those attributes that he/she wants to add and produce a
> cumulative one via simple transform/document function. This is of course
> more robust then to modify the original.
>
> 4d. Providing/changing binding for the test harness. If harness would
> require a change in the metadata, produce/supply the needed metadata in
> a separate file. Provide the original copy.

Yes, this is very useful. Historically we've let the WGs choose which 
license they want to use given their own discussion of these issues, and 
now with the further insight from this discussion we can make it clear that 
the harness/binding could be licensed differently than the test cases. 
Furthemore, the W3C isn't litigious as our goals are "wide distribution of 
our work, absent confusion and misrepresentation." We're obviously not 
going to get upset by someone acting in good faith who made a tweak to a 
test suite because of a bug, and noted/reported the bug. But if someone was 
maliciously misrepresenting the state of conformance and causing confusion 
about our specifications we can exercises our rights to remedy this.
Received on Wednesday, 26 February 2003 10:09:38 GMT

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