[Insight-developers] A Contrib group?

[Brad King]

On 01/16/2014 04:31 PM, Bradley Lowekamp wrote:
> It's was not clear to me that a separate group for contributions
> was proven not to work. Certainly referring to it despairingly as
> "second-class" is not inspiring. Specifically what were the problems,
> from the previous Review?

As of ITK 3.20.1 there were almost 350 classes in "Review":

 $ git ls-tree -r v3.20.1 -- Code/Review |grep '.h$' |wc -l
 349

A manual check of all commits ever to remove files up to that point:

 $ gitk v3.20.1 --diff-filter=D -M -- Code/Review

shows that only about 20 classes were ever moved to other kits.
Only a dedicated sweeping effort during modularization moved them
elsewhere, and that still left behind about 70 classes as of 4.5.

> Additionally, How is steering new contributions into remote modules,
> not creating a tiered system?

I argued against using remote modules for new contributors or review
and argued in favor of direct review in Gerrit for upstream integration.

-Brad K