In the late 1990s, the physics department at the Massachusetts Institute of Technology (MIT) had a problem. The department was responsible for teaching the two required physics courses that are part of the General Institute Requirements (GIRs), MIT's core curriculum--Physics I (mechanics, or in MIT parlance, 8.01) and Physics II (electricity and magnetism, 8.02)--and the failure rate in both was dismal. Often as many as 15 percent of the students didn't pass mechanics on their first try, and 10 percent didn't get through electricity and magnetism. The department head, Marc Kastner (now dean of the School of Science at MIT), and the associate department head for education, Thomas Greytak, were under pressure from the senior administration, as well as faculty in other departments, to fix the problem. This is a story about the transformation of the two foundational physics courses at MIT from lecture/recitation to the kind of interactive teaching and learning that educational reformers hold up as a model of best practices. Named TEAL--Technology Enabled Active Learning--the format is still used in Physics I and II at MIT today. The account of how TEAL was born and managed to survive demonstrates just how difficult substantial educational reform can be. TEAL's story is especially important because calls for reform in STEM teaching have been particularly urgent and persistent, with long-time advocates observing that the most pressing need has been for improvement in introductory science courses, because they can make or break a student's interest in science, mathematics, or engineering. TEAL is a textbook case for what it takes to reform such courses. The ingredients for successful pedagogical change were present (departmental and institutional support, a faculty champion, funding), as were some typical roadblocks (critical faculty, conservative students, a reformer's burnout). As is true of most stories, TEAL's has its protagonists and antagonists, its high and low points, its challenges and victories. Whether it will have a happy ending remains to be seen--in part because the very definition of what would be a happy ending depends upon whom one asks. (Contains 14 resources.)