Let's capture the really important point first: we should not create 'educational versions' of students' favorite applications. Not simply because, as Catherine Howell suggests, they will not challenge students. But also because, as Cole comments, students wouldn't use them (at least, not in the same way). Because "they said things the U runs feel 'controlled.'" There is, as Ulises Mejias suggests, a connection between the software we choose and the social structures we set up. That's why I suggest we should be looking for specific properties in the software we select. What does this mean for educational software then? That we should not be trying to build separate systems, but rather, we should be building object and content types that integrate to, and become a part of, their already existing environments today and into the future.