Process, etc.

[Havoc Pennington <hp redhat com>]

Miguel de Icaza <miguel ximian com> writes: 
> I did that because I found a few mails that began with `Fuck you
> Michael' and I saw a kind of debate that was far from being
> rational. 

Even now with a cooler head I won't apologize for that. I simply will
not respond politely to any mail containing conspiracy theories about
the "GTK Team" or "Red Hat" instead of assuming I'm an individual with
equal standing in the GNOME Project. Sorry. Maybe he should try
rephrasing whatever he was trying to say.

In the meantime, I hope we can have a nice conversation with you and
me and Dietmar and other interested parties.

I'm hoping you will join Daniel and Dietmar and others in addressing
the question I've asked about how we could have avoided this. I
politely discussed this bonobo-config/GConf issue with you and Dietmar
some months ago, with Colm etc. participating. If you go back to that
thread I thought it was technical in content, quite rational, with no
use of the term "fuck" IIRC.

However it was apparently insufficiently confrontational to get any
results, and we ended up postponing the issue until it had to be
brought up again in a situation closer to deadlines and with people
more strongly invested in their code. Thus this heated flamewar.

So my question is, what could we have done back then to avoid this
situation? I don't think "just be nice and loving" is an answer; I
don't think "just agree with us" is an answer. What would an answer
have been?

My view is that the answer was that thread had to continue, possibly
accompanied by phone calls or in-person meetings, until we actually
all understood the issues and arrived at a common
conclusion. i.e. agreeing to disagree was not a viable option, we were
deluding ourselves. Do you think that's right?

Now, the value of a process as suggested by Daniel is that it would
have forced some kind of official decision, which would have forced
that thread to continue because we couldn't both go off and pretend we
were going to be the "chosen config solution," a decision would
actually have been made at that time.

What is your alternative suggestion for ensuring that people really
confront issues and solve them together, rather than sulking off to do
their own stuff?

And moving beyond that old thread, what is your suggestion for the
present as to how we make this decision? Or are we currently in
agreement? I think we may mostly agree.

Elliot has been talking to Chris Lahey and is mailing Dietmar to
collect technical information; he's thinking of posting a summary of
the technical issues from a third-party viewpoint, starting with
requirements and moving to how the various solutions meet them.

As he suggests I think if we aren't in technical agreement we should
first debate the requirements then sort out how to achieve them via
the GConf or bonobo-config codebases.

But the question still remains, if we have this technical discussion
(again), and still don't agree at the end, how do we address that?
We do have to make this decision.

I'm hoping you have some insight into this issue.

Havoc