Nova Spivack has a short blurb titled Minding the Planet: Communities of Purpose: The Third Type of Community. He mentions wanting to create tools to help people create more productive communities of purpose.
The question then, is what should such a tool support? The problem is that the answer is "it depends". It all depends on the purpose, which I guess is why most existing tools focus on what Spivack calls "communities of interest"
and "communities of practice".
It's "easy" to create a generic tool like Yahoo! Groups that supports generic groups with discussions.
However once you move to communities of purpose, tools are likely to need to be more specialized. Consider Groklaw, and how it could benefit from tools allowing better support for managing the plethora of legal documents they host.
SourceForge
is perhaps the best example of a tool designed for a community of purpose that I know of. It's a tool designed to help support the purpose of managing software projects, allowing you to easily create a custom community around your project.
The problem of course is that it's a niche tool, and while SourceForge's niche is quite large, it's easy to imagine a huge number of purposes that would benefit from specialized tools, but that are too narrow niches to warrant building a tool from scratch.
How do you blend existing types of tools to create something that is easy enough to customize without software engineering experience?