Alternate Title: "How to start a collaborative writing project or die trying,"
Step 1: Lock yourself in your office, fire up Emacs, and write an initial draft from beginning to end yourself.
Step 2: Post it on the Internet.
Step 3: Encourage contributions and hands on feedback from your collaborators. Have a piece of cake.
You may thing that I'm kidding, but it's true. I think there's a misconception that the way to write something with other people follows a path that might look like: having a meeting to establish the common goals and an outline of what needs to get said, and then another meeting to divide up who is going to write want, and then people go back and write their little parts, and then you mash them all up and everyone rewrites their part till it meshes with the other parts, and then you pray it says what you need it to say, and doesn't need further revising--except it sort of dose, so you repeat the whole process over again, to revise the text, except with editing instead of writing. And because each stage requires endless conversation, when you don't have the benefit of face-to-face meetings, things can take a long time: so long, in fact, that most people will have probably lost interest long before something has been written. The short of it is that this method, though very democratic and open seeming, isn't.
I think there's a fear, that when a single person puts a lot of individual energy into a text (or any kind of project, really,) and doesn't consult with collaborators at every turn that it somehow becomes not a collaboration. This is emphatically not true. There is significant difference between endless group process and the collection of meaningful feedback; a real distinction between a text created with a process that involves many people, and a text that many people can agree represents their interest, purposes, and needs.
I think, though I'm not certain, that one could replace the words "writing" and "text" in the above, with "programing" and "code" but I don't know for sure.
At work, I have a moto: "you can't edit it if it doesn't exist yet." The more interesting thing, I think, in every context is when you go off into your own office, fire up the emacs, write something, and then say "so how does this look?" People sometimes say, "nice, but you used 'setup' as a verb in the third paragraph," or "ok, but you left out a section about flux capacitors in section two, and I think that's crucial for understanding most of section three," but these are problems that are fairly easily addressed.
Now it could be the case that I'm just that awesome (unlikely), but I think it boils down to the fact that most people don't understand how to make texts. I also think that a lot of "group process," can be obsoleted by an individual who can produce something, and has a good sense of the group's desires, and who knows how to check in with various group members at the right moments. While these skills can be listed quite effectively, and it's true that there is no rocket science involved: some things are easier said than done.
Not every collaboration works, and there are a lot of variables at play in any situation where a group of people must come together to make something, but in nearly every situation beginning with "hey, I want make something with you, look at this draft," is better than "I was thinking about making something with you but I wanted to get your feedback first."