RE: QF Policy & Op guidelines

Hi, Andrew!
Thanks for raising this, we discussed these 3 points on today's telecon,
outcome is inline (WG members, please check if I correctly summarized
this):
 
>   1.
>   Test material development. Do we have a global license for use and
>   distribution of test materials. Or is it per-group?
[KG] We currently address this in the checkpoint 6.3 and 7.3:

Checkpoint 6.3.Priority 2.In QA Process document, define the licenses
applicable to submitted and published test materials.
Note: Working Group can choose either the W3C Document or the W3C
Software license. If applicable, W3C Document license is recommended

Checkpoint 7.3.Priority 1.If plan to transfer the test materials to W3C,
resolve IPR questions and reach agreement with external party that
produced test materials.

We agreed to link 7.3 from the note for 6.3, as there might be other
licenses that the WG has to sort out with external party in case if the
test materials were developed outside of the W3C.


>   2.
>   Guideline 3. "...developed test materials can be used by external
>   parties including certification services." We Need to explore the
>   liability issues here.

[KG] We ask WG to put the following disclaimer in Checkpoint 3.3:
Checkpoint 3.3.Priority 2.Provide disclaimer regarding use of the test
materials for compliance verification.
....
1. passing all of the tests does not guarantee full compliance of
implementation to specification 
2. failing test suite means failing tests for specific feature they
target
 
In fact this disclaimer raised many discussions, specifically the part
2. We decided to leave it as is for FPWD, in the mean time comments are
welcome on the list.
Argumentation for the current 2.:
  - Existence: Disclaimer needs to be complete - if we give definition
of what does success mean, we need to give definition of what the
failure mean
  - Wording: Objections were raised by Lynne and Dimitris that failing
test means non-conformance. But everybody agreed that the term
"non-conformance" is almost always conditional, leveled, etc. Apart of
this "conditional" nature, formula "failed - means non-conformant"
doesn't apply for tests for discretional behaviors - this was already
pointed out in the current OASIS XSLT committee work. 


> 
>   regards,
>   Andrew

Received on Tuesday, 29 January 2002 02:04:36 UTC