Although collaborations are large, the work gets done in small groups. Different groups are responsible for different pieces. Nobody knows every detail of the entire experiment. Wouldn"t it be cool if we could have a $1B proton machine in Urbana? There are more thesis topics than there are graduate students by a lot! Hardware development projects in Urbana are typically done by 2-4 people. Data analyses are typically performed by groups of 2-4 people. What Do Students Do?