"David Abrahams" <dave@boost-consulting.com> wrote in message news:upt6ibdyw.fsf@boost-consulting.com...
It's up to you whether to introduce yet another kind of review process, but I'm concerned that might end up with a hodgepodge of not-very-thoroughly-inspected and not-very-interoperable components. So, how will we keep the overall quality high?
I propose that we use some time to establish a reasonable amount of categories of algorithms. We already have string related, I'm working on statistics related, and there are probably more categories out there. Then, once we have these categories, one or two persons (preferable persons with domain knowlegde) must take responsibility for the overall quality of that library and inspect new mini-submissions and work with the new authors to ensure their contribution will fit into the overall scheme. Then a formal mini-review should follow. br Thorsten