Patch Pilot report

[Andrew Cowie]

On Tue, 2009-11-24 at 16:44 +1100, Andrew Bennetts wrote:
> I'm more uncertain about the long term effects... will this mean we
> handhold contributors indefinitely when they might have graduated to
> self-sufficient sooner?  My guess is no; a sufficiently motivated
> contributor will learn from the example of the tests I write for them,
> and try to do that themselves on the next patch, etc. 

I agree with this sentiment.

My experience, however has been contrary to it.

Again and again whenever I have just done it for them and pushed a patch
over the edge and even tutored a contributor on what "would be nice", I
consistently get exactly the same garbage when they submit their next
contribution.

Indeed, I have had comments like "huh? It's your job [maintainer] to
clean up (radically fix, realign with project architecture, correct
style, write documentation, write patches) patches".

[I believe this to be a legacy of the Apache worldview of accept
anything + someone else can fix it if they want to; reality is that this
doesn't actually happen and code quality deteriorates in non-core areas]

Despite my best efforts, it's only been rare that I've managed to get a
contributor to raise their game, and it's never happened when I just
fixed things for someone.

Which is frustrating, because if I play the ogre and keep the bar high
and push back, most people just go away.

I think that in our case, part of the learning curve is DVCS itself;
people don't know to look at what I did to their branch before merging,
nor how to diff against ancestors, etc. The fact that all of this is
doable with bzr doesn't change the fact that it's non-trivial even if
you do know such things are possible.

Anyway, I think you guys have the right approach with your patch-pilot;
bzr's barrier to entry is very high, and it's worth helping people get
over the line; I think it's also pretty clear that this is a one-or-two
time service, not automatically a given for all future contributions.

Despite the bad taste in my mouth, I remain an optimist; I still think
it's worth trying to get new contributors to learn from the experience.
Perhaps if your code review system was to respond with: "your patch was
merged, but this&this&this (ie, diff) were changed from your submission
to acceptance..." it might help with the learning factor. Hard to say.

AfC
Sydney


-- 
Andrew Frederick Cowie
http://www.andrewcowie.com/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20091125/2d2dedc3/attachment.pgp