[development] #drupal and #drupal-contribute split (Was: Re: Proposal: Move all dev support off this list to new StackExchange site)

Earl Miles merlin at logrus.com
Sun Mar 20 17:11:22 UTC 2011


On 3/20/2011 10:00 AM, jeff at ayendesigns.com wrote:
> I agree. I would have loved to contribute directly to D7. I found a
> level of...discomfort overall. I've was coding when many of you were
> first learning to count, so it's not the php I'm uncomfortable with.
> 
> 
> I think part of it is coming upon a group that have been together so
> long that they're clique-ish, not meaning to be and not in a bad way,
> but they know their routine and topic and each other so well that
> everything is in shorthand, and it's great for being a voyeur but
> daunting to a new guy. The other part is some discomfort with the
> fast-paced immediacy of #drupal-contribute when combined with being a
> newbie to core, like jumping on a fast ride that's already moving.
> 
> 
> I would have felt more secure about it even with something as small as
> some color coding on open issues that signified "this one is probably
> good for a core newbie."

I suppose there is some clique-ish nature to that, but it's really more
about level of trust.

It's fully possible to get into the trusted group.

That said, it's not just about that. I find core development difficult.
I'm in the group. But it's a long, tedious process and you need
excellent debating skills and a lot of time that isn't spent all at
once, but is spent in dribs and drabs over the life of a patch. The more
complex the patch, the longer it will take.

Getting a 1 line change to core is pretty easy.

Changing 500 lines is pretty hard.

Our review process is incredibly difficult; reviewing is boring, and
tedious, and not rewarding. Not many people do it. Those who do it are
reviewing code primarily for style, because that's the easiest thing to
review. Doing a good code review requires understanding the bigger
picture, and there isn't a bigger picture anymore. Maybe there was once,
but Drupal has lost it, because there's no one who actually knows all of
Drupal anymore. The bigger picture is an organic mess of Chaos and each
reviewer has his or her own internal "bigger picture" that is reviewed to.

Sometimes the problem is just getting a reviewer. Sometimes the problem
is getting past a reviewer's pet peeves. Either way, the personal cost
to code for core is high, has always been high, and only gets higher. I
don't really see how that can change. Ideally we want to ensure quality
contributions, and and still be receptive to all developers, and those
two policies are at odds.


More information about the development mailing list