Collaborative book production can take many forms. It can be a managed closed group - similar to traditional publishing - or it can be a more open flow of contributors - similar to wiki production models.
Both have their merits but by far the most popular is the managed closed process. The managed process can be extremely tedious and slow. For a good account of how tedious you may wish to read more about the OReilly process as it is outlined online1. The Oreilly process is lightweight in comparison to some (notably academic) book production processes but it's bad enough. It is because of the tedium and slowness of these processes that I have avoided them and instead I have been unpeeling those managed layers and peering into the supposed abyss of open collaborative book production for the past 5 years. The model is simple - anyone can jump in and change any book at anytime.
This approach is not common however. There seems often to be an unspoken assumption that control is necessary and books must be protected. Protected from harm – not just the malicious kind but harm inflicted by contributions that lower the quality of the text. My experience from 5 years running an entirely open system (FLOSS Manuals) is that there is little to fear and much to gain by losing control. In five years running FLOSS Manuals I have not seen a single malicious edit. It seems to be the case that if people are not interested in your book they will leave you alone. I have found that the approaches to the text are sensitive and respectful and more often than not they improve the work – sometimes in very surprising ways. If you open yourself up to collaboration it is very interesting what can happen. On one early open book I worked on a retired copy editor who I had never met (at the time) went from top to bottom of the 45,000 word text in his free afternoons and made an incredible improvement to the text. I would like to have thanked him at the time but I didn't know who it was until later.
The trick is not to protect the text but to manage the social processes surrounding it. It is actually hard to get people to collaborate - most people do not want to disappear into the authorial reputation of someone else - which one reason why equal credit for any contribution is important. You need to work on coaxing people to get involved and some are harder to convince than others. The best strategy to get people involved however is to make it as easy as possible. Do what you can to welcome people in and make the process easy (both technically and socially) - it will be harder than you think to get uptake but the results are rewarding. The process is very much like open source software production. In early waves of open source mania many projects were started with the belief that the coders would 'just show up'. But communities (most anyway) do not develop like that. They need someone dedicated community leader / manager / guru to put the time in and make people feel part of the process. A book is a community and it needs this kind of leadership - a community of book producers with multiple books needs this even more.
Unlike open source software development however open book development benefits from relinquishing as much control as possible. The best philosophy is to make contributions easy to make and then manage the result. Open Source usually works the other way round - vet contributions closely and avoid calamity. But in my experience this does not work in the book world and it doesn't need to. Books can be vetted after the contributions - nothing gets broken if there is a mistake (unlike code) and so put in place technical and social processes to watch the content and manage the result.
It is often harder than you think to attract attention and contributions. It often relies on how effectively you can get the word out and how attractive you make the offer. You need to reach out to people and inspire them. The more direct the approach the better – one to one face to face conversations work best, personal emails work well, emphasising concrete outcomes is very likely to improve results, as is making the offer fun, relevant and illustrating a real need. But the usual rules apply for attracting collaboration in any realm – its a mix of luck and getting the tone and channels right.
Once the contributions start rolling in then it’s up to you to manage this process. If things get out of control clone (copy) the book and decide on a more moderate development approach. However the best tool for managing input and getting the book to where you want it to be is social management. You need to coerce the contributors to come along with you and share your vision of what the book should be. At the same time you need to also be able to make the process satisfying to them which means that the contributors must also have a say on what the book should be. The process needs to be a moderation of both top down and ground up processes.
So ‘how to control’ a book is a question with some merit but can only be answered from the perspective of managing social processes not managing content. If you can lose the feeling that you must control the book and instead relinquish as much control as possible you will be surprised and very probably excited by the results. In a world of free culture its all about the art of losing control…