[ODFPlugtest] Chatham House Rule

[robert_weir at us.ibm.com]

It might make sense to craft some language to state exactly what we agree 
on for protocol.  Chat House Rules I think expresses the "spirit" of the 
desired rules of engagement.  But it is not 100% directly applicable, 
since it is more concerned about the individuals and not companies or 
products.

Didn't we have some basic "principles" from the Hague plugfest?  Maybe we 
add something to that?

-Rob

plugtest-bounces at opendocsociety.org wrote on 11/20/2009 09:06:34 AM:


> 
> Rob,
> 
> Should I say anything hopefully helpful?  
> 
> Basil

> 2009/11/20 <robert_weir at us.ibm.com>
> plugtest-bounces at opendocsociety.org wrote on 11/19/2009 06:50:15 PM:
> 
> > Personally, I think these rules create a complex situation of that
> requires
> > people to keep a duality in place about what can be said and what not
> and
> > can you infer things I cannot say from things I am allowed to say. 
Also,
> > since many bugtrackers are public, Chatham rules would stop bugs from
> being
> > allowed to be reported.
> >
> > Being cooperative is an important goal and 'Chatham rules' serves as a
> > synonym for 'be constructive' and I appreciate it as such. Strictly
> > following the rules seems impractical though. In my mind, what is more
> > important is that the standardization process allows a low threshold 
for
> > new implementors.
> >
> 

> That is a fair point.  One of the primary goals of the plugfest is to
> identify interoperability bugs between implementations.  I'd fully 
expect
> that Microsoft or IBM or Google would take what bugs it found and report
> them to their development teams.  So at IBM we might enter a bug that 
says
> "Symphony 1.3 does not process feature X from implementation Y" and if
> this was due to a bug in implementation Y we would say so in our defect
> report, and then discuss how to work around the bug to the customers'
> benefit.  I'm sure something similar would happen for other proprietary
> products.
> 
> Where it gets interesting is when the ODF implementation is open source
> and has its defect tracking system public.  Surely we want open source
> implementations to be able to leave the plugfest with a set of bugs to
> look at?
> 
> So I don't think we want to be so strict as to prevent all implementers
> present to make use, as engineers, of the information they received in 
the
> plugfest.
> 
> But what we want to avoid is things like:
> 
> * Ascribing a statement to a specific person or company.  We shouldn't 
be
> reporting, "Doug Mahugh from Microsoft said that Office 2007 had bugs X, 
Y
> and Z".
> 
> * Speaking derogatorily of another product's performance at the 
plugfest,
> especially where pre-release software is shown.  We want to encourage
> vendors to show their beta and earlier software.
> 
> Of course facts that are know from outside the plugfest are fair game.
> Otherwise I'd just come in, show a list of all known Symphony bugs on 
the
> screen for 5 minutes and declare that no one can ever talk publicly 
about
> Symphony bugs in the future because they are all covered by Chatham 
House
> Rules.   The point is the confidentiality applies to the activities and
> statements of participants in the plugfest.  But if you later find the
> same bug in the publicly available version of Symphony, then I don't 
think
> I have any expectation that this public fact will remain secret.  Of
> course, you still would not want to discuss conversations about it from
> the plugfest or ascribe statements to participants at the plugfest.
> 
> The overarching theme is you want to make it safe for engineers to
> participate in the plugfest and show, test and discuss code that is not
> yet perfect.  But facts concerning publicly available code -- I don't
> think we can expect secrecy about those.
> 
> Regards,
> 
> -Rob
> _______________________________________________
> Plugtest mailing list
> Plugtest at opendocsociety.org
> http://lists.opendocsociety.org/mailman/listinfo/plugtest
> _______________________________________________
> Plugtest mailing list
> Plugtest at opendocsociety.org
> http://lists.opendocsociety.org/mailman/listinfo/plugtest