I'm a qualified Prince 2 Project Manager, and my comment is this...you cannot run this like a normal project, it WILL fail. There are too many uncontrollable risks (commitment, decision-making processes, completing work packages to a timescale, etc, etc, etc)
The first thing you need for a community project is enthusiastic participants, with the necessary skills but more importantly, the necessary time. You need people who are prepared to compromise, and don't always want their own way. You need people who will take their task and complete it to the specification, without deciding they have a better way to do it. You need a small team of people who connect the dots, rather than make the dots. And you need a project steering committee who lead the project for the good of the project, not their own egos, and can manage a distributed team of developers with every quirky personality under the sun. The steering committee can include the individual with the big idea, so long as that person can let go of their personal attachment to the idea and allow it to be nurtured by a surrogate family.
Having said all that, it's achievable once you lay down the rules and everyone embraces them for the better good. I don't fit the bill, because I am lacking in time. Make sure everyone declares their commitment, and the time they have to throw at the project. Make sure they all want to be part of an idea that isn't necessarily their favourite idea. Ensure you have fulfilled all of the roles and ideally doubled up to account for dropouts. Then you can declare the project initiated and ready to start.