As Seth Godin points out in his post today, the success of a joint venture is greatly reduced due to both:
The Venture Part: Is by definition risky.
The Joint Part: The more independant parts in something that can fail, the more likely the overall thing will fail. This applies to just about anything and gets even more challenging to predict when dealing with the complexity of people.
Open Source Software Projects
Open source software projects are often done by a team of dispersed and somewhat independent developers, etc. The common goal keeps everyone aligned.
If someone volunteering their time drops out, someone else in the community must step in and fill their shoes. If the group is large enough, losing one or more people poses little risk. The downside is that these projects can take quite a while and timelines are often variable.
The problem with the comparison is that open source is not for pay, it is volunteer work by definition. This makes sure that only people who are truly motivated for the project to succeed will show up. But I don't think open source can be categorized as a venture in a business sense if there is no compensation in the model.
Goals Alignment and Tracking
The final piece is making sure the goals are aligned and tracked. There is something to learn there in open source projects.
They generally use tracking software, transparency and a lot of peer pressure to deal with underperformers on the team… this is the accountability piece and in the development world it can be powerful.
For your joint venture you need leverage (something that works in your case) on enforcing commitments and a workable plan for dealing with partner breakup while having the venture survive.
Or you can take Seth's advice and keep the venture under one roof.