Start with dinner. Get together. Talk. Dream. Learn. Over food, of course. Don't be self-centered. You'll kill a relationship quickly if you spend all evening talking about yourself, your needs, your wants. Instead, find out about your potential partner. Learn about their hopes and dreams. Think about how they may enhance or build on yours. Don't name the baby. Put off the discussion of branding, naming the project, how credit is bestowed. This gets emotional fast, and quickly moves you out of the realm of low-risk prototyping. Put off the prenup. In fact, I'd avoid writing anything down at first -- especially anything regarding goals, directions, duties, etc. This starts to define the relationship from the outset instead of allowing for open innovation and low-risk experimentation. Respect each other. Be nice. Be giving. Be open. And if that costs a little, consider it an investment in the potential of the partnership. Pick up the check here and there. Meet up again. And again. Make a plan -- and put it in your calendar -- for the key people to meet regularly, preferably over a meal, to check in on how everyone's doing. That's the time to make sure nobody feels disrespected, over-committed, or unhappy. Then adjust accordingly. Break up gracefully. If the partnership just doesn't click, part ways, remain friends, and be sure your team gets together to learn from, and record, what parts worked.I won't kiss-and-tell about our newest collaboration, but I will say this is the approach WNYC took when we approached Iowa Public Radio back before the Iowa caucuses. We made a concerted effort to learn about them and focus on their needs. We talked a lot. We shared info and a common effort. And we didn't name the baby. The result was an amazing night of radio, and smiles all around (scroll to the bottom). It's also how we've approached a lasting relationship with the wonderful folks over at the Connecticut Public Broadcasting Network, where we first prototyped this kind of coverage.Happy dating![Photo by hypertypos]