We have multiple layers of governance in place. The Community manager is the first line of governance, most of the time I can resolve any issues myself. Not looking forward to those kinds of problem on any more. Sense I left my old community to NetApp I have not had to ban one person.

We are in the thick of adoption on a year old wiki/blogging implementation to 17,000 and as one of the two founders and project managers, I fought tooth and nail every time any one brought up the word Governance. We have been extremely successful so far by not forcing a rigid framework for usage, but instead leading by example. We find users who inherently understand the "proper" way to interact with corporate social media and encourage them to take on more responsibility. I suppose an organic governance model sort of emerges from that, but it is at the mercy of the users. Not IT, and definitely not me. The project was conceived by users, is administered by users, and users make up 99% of the rules. Sounds like a bit of a pipe dream, but it is possible if you focus on enablement and riding the wave wherever it may go, as opposed to governance.

True, it is an internal implementation. I tend to focus primarily on internal implementations for my clients and my company. I have a few thoughts on some of your questions though. Again these are for internal implementations...

Policy Setting - Went with the "Use Your Head" policy. There was some early discussion about stating which system (Doc Management, Intranet, Wiki) folks should use for various purposes. We scoffed stating that policy should not be a veiled attempt at tool bigotry. The usage matrix ended up being made but adherence is limited. Offer compelling tools and users will gravitate toward the ones they find easiest to use. Oh...and make sure you have a good search engine :)

Functional Group Turf Wars - We are a professional services firm. I think turf wars are SOP. We encourage them to happen in the open so that the wisdom of the crowds can determine who is right and who is greedy. More often than not, one side is not willing to do battle in the clear. The other side de facto wins. 

Multi-level governance model:

-- 0th level: Terms of Usage posted with some very simple "guidelines" (not policy, not rules, etc.)

-- 1st level: community managers and/or helpful individuals

-- 2nd level: corporate platform managers

-- 3rd level: exec sponsor (me)

-- 4th level: ad-hoc committee of exec VPs (IT, HR, etc.) if we encounter something that requires some serious discussion

Levels 0 through 2 are open and transparent, e.g. anyone can comment and/or contribute.

Since Sept 07, we've only had a small handful of issues that haven't been solved at levels 0 through 2 -- very minor, just growing pains.

We have an E-governance Board.  They consist of representatives from major stakeholder organizations (agencies).  Most of them work in their organization's communications department.  They make the high level decisions and provide guidance to the Clearspace implementation and the external facing website.

I'm the E-government manager who serves as staff to that committee along with my staff.

Each Agency has a high level space and Community Manager and sub community managers.  We push down as much control as the Clearspace product will allow us to so they can support their space locally.

We have a support space that my group monitors to help the Community Managers and regular collaborators.

We have created a policy (ver. 1.0) around use for those who participate.