[development] Subsystem maintainers or targeted committers?
Larry Garfield
larry at garfieldtech.com
Mon Jul 2 15:06:05 UTC 2007
On Mon, 2 Jul 2007 07:37:29 -0700 (PDT), Karen Stevenson <karen at elderweb.com> wrote:
> I think this is starting to get somewhere :-)
>
> I see the biggest problem is that there is nothing now between 'Feature
> Request' and 'Patch to be Reviewed'. In fact, it seems that you can't get
> much attention to a FR unless you provide a patch to be discussed, which
> puts the cart before the horse and probably encourages everyone to just
> jump in and start coding without enough support around the idea. Perhaps
> we need to move Feature Requests to a different system or different part
> of the system where you can propose a feature and have a forum for
> discussing in a global way whether the feature is needed and what should
> be included in a solution. Then, after discussion has solidified and there
> is agreement that it is a good idea and what direction it should take, and
> only then, someone creates a patch to be reviewed that is put into the
> issue queue. And any non-trivial core patch should be somehow be forced
> through that system before it will be considered.
>
> Now, what needs to be done to create a system like that? I'm not sure.
>
> Karen
The simpler version would be a new status "concept needs feedback" (or something like that). Indicating "please talk before I write code so that I can write code." For most non-trivial changes that's probably a good idea anyway (design first!), and would be an explicit flag for subsystem experts (formal or informal) as well as committers to jump in and nip a dumb idea in the bud, correct a design flaw before it gets codified, or indicate that they think something is a really good idea.
And it keeps it on-site rather than on people's blogs. :-) It could eve have its own custom link (like the "patches needing review" link) so people can see what's being discussed at-a-glance.
--Larry Garfield
More information about the development
mailing list