[ANNOUNCE] dh_splitpackage 0.1

Daniel Holbach daniel.holbach at ubuntu.com
Sat Jun 4 19:20:55 UTC 2011


first of all let me say that we agree on a lot more than we disagree on.

Am 04.06.2011 20:20, schrieb Scott Kitterman:
> I'm quite encouraged by Zygmunt's latest reply in the thread.  It's more like 
> the reply I would have hoped came in response to Evan's mail.

In an ideal world every single contribution would go upstream first, be
general and solve a lot of problems. Failing that, after some
conversation, the patch author and others would agree on a good way to
get the general solution into place. I agree with all of that.

I just don't feel the way this whole conversation went was the best way
to get there. There was somebody who put a lot of work into solving a
problem and shared the solution. Where we might disagree is if it's a
reasonable to expect the initial patch author to put even more work into
it. It would be great if it happened that way, but it will cause
frustration if that expectation isn't met.

If we want that person to put even more work into the solution, I feel
the discussion needs to take a different route. Admittedly there was
some confusion about email addresses which didn't help, but still I felt
that in this particular case Zygmunt was dismissed in much too rough a way.

>> In a lot of cases in the open source world somebody proposed a solution
>> to part of a bigger problem and even if they chose not to completely
>> generalise it, upstream it, etc., it helped others to pave the way for a
>> more general solution. It would be great if all proposed changes in the
>> world landed upstream first and in a general way, but I don't think it's
>> a fair a priori expectation.
> I think you are putting words in my mouth now.

The only way I could read
  "You don't appear to have any interest in collaboration to improve
   things for the greater benefit."
was as surprisingly general and as a dismissal. In addition to that
  "Fortunately your response to his helpful suggestion has reduced the
   risk you'll ever be inconvenineced by such responses again."
to me came across as quite cynical and indicating that if Zygmunt was
not willing to generalise his solution, it'd be pointless and he'd be
better off not contributing it at all. Certainly it's not welcoming the

If I misunderstood you, I'm sorry. I don't want to dwell on the question
of particular words for too long as it does not help moving this
particular issue forward, but a different tack would have been:

 - Discussing with Zygmunt what particularly could be generalised most
   easily and asking for his help if possible.
 - Asking if he'd mind if Upstream (and others) would be copied in the
   conversation or if the tool could be submitted for a code/concept
 - Interpreting the tool as a proof-of-concept which could be
   generalised later on.

The reason why I'm imagining a softer approach to drive the discussion
is that I have sympathy for somebody who just solved their own problem
and shared the idea, but might be too busy / to much out-of-the-loop to
figure out next steps or actually implement it.

This feels a bit like sponsoring, where somebody sent a patch that
solves their problem, but not the complete surrounding problem and it's
not submitted upstream first. The question is: do we accept the work and
gently help moving things forward or do we expect the contributor to go
all the way and ignore the solution until then.

>> This exchange does not only alienate Zygmunt, but also future
>> contributors who happen to read this. Everybody is entitled to be of the
>> opinion that solutions are worthless if they don't generally fix all the
>> related bugs, but you don't speak for me.
> That isn't what I said at all.  Personally I find this kind of response to 
> attack someone who's calling someone on unacceptable behavior demotivating.

What exactly was unacceptable? Could you imagine that Zygmunt was
demotivated as well?

>> Ubuntu is different because we invite people to share their ideas and we
>> welcome people in. Dismissing a helpful developer is unproductive and
>> more importantly actively damaging to the project.
> I'm not sure who the helpful developer you're referring to is?  Blasting 
> someone who offers suggestions about how best to get one's work incorporated 
> into Ubuntu is not, IMO, helpful.  

I'm not sure Zygmunt's intent was to get dh_splitpackage into Ubuntu. At
least it's not mentioned in the initial email. The way I read the mail
is "Hey, I solved a problem I ran into, check it out if you're interested."

> I see in another part of this thread that 
> Zygmunt is going to work with the Debhelper upstream to see if this can be 
> incorporated.  I think that's very good news.

I agree that it's good news, the question is if this was the most
pain-free way to "get there".

> I think that sitting idly by while people are hostile and
> negative makes the environment more difficult for everyone.  

I totally agree with this and I did feel the need to step in.

> Instead of a long 
> rant attacking Ubuntu and Debian, Zygmunt could have just said he wasn't 
> interested in doing that work, but he didn't.  He went on the attack and I 
> think it's unreasonable for you to attack me for calling him on it.

There certainly was frustration in that mail, but it also contained
valuable feedback about how it feels contributing something to
Ubuntu/Debian for somebody who is not participating in Ubuntu
development every single day.

I think it's absolutely possible to point out that the tone of an email
is not OK, the critical difference between our two perceptions is in
"rant of a frustrated contributor" vs. "attack of our projects".

>> Negative:
>>  - some confusion about email addresses,
>>  - a lasting impression that contributing to Debian and Ubuntu is hard
>>    and you might get flamed if you share your work but might be too
>>    busy to fully generalise it, etc etc.
>    - Ubuntu developers getting flamed for upholding project values

It seems we both clashed in trying to do the same.

>    - At least one Ubuntu developer feeling like the Canonical community team
>      is more committed to Canonical employees than Ubuntu.

Jono, Jorge, David and Ahmed were not part of this debate.

It's an idle side-path of the discussion, but I don't want anybody
treated on this mailing list like Zygmunt was treated, Canonical
employee or not.

>> Can we please go back to square 1 and can those who are interested in
>> the problems that are solved (or partly solved) by Zygmunt have a
>> conversation that is goal-oriented instead? I'm sure Zygmunt is happy to
>> answer questions about how his code works and which considerations
>> exactly led to it.
> I'm happy to take a step back and consider a better path forward.  I think 
> Zygmunt's reply to maco is very encouraging.

I agree that things it's great we have paths of communication
established now. I just feel that we could have got there in a much
calmer way than we did.

Again, if I misread any of the previous conversation and misinterpreted
it, then I'm sorry. There was other confusion part of this debate as
well, which probably also didn't help.

Have a great day,

Get involved with Ubuntu Development: http://identi.ca/ubuntudev
http://twitter.com/ubuntudev http://facebook.com/ubuntudev

More information about the ubuntu-devel mailing list