Wm. Shakespeare said that.
Todd Murphey (’93) said this:
This video is a (somewhat biased) comparison between simulations created with a variational integrator (left) and a continuous ODE (right). The variational integrator is a Mathematica implemenation that ran in about 10 minutes with a 0.01 second stepsize. The continuous system used a C implementation that ran in about 30 minutes with a 0.0001 stepsize. Despite the advantages, the ODE-based simulation does not handle the string constraints properly and blows up after about 23 seconds. The variational integrator, on the other hand, avoids any energy misbehavior and leads to the desired simulation.
We so wish we understood. But we majored in English and took our last math class in 1993 CE. Thus, we’re limited to reading the Bard and watching Todd’s cool video on a continuous loop:
There’s more exciting stuff on the website of the Automated Marionette Project, one of Todd’s undertakings with colleagues at the University of Colorado, Boulder, where he’s an assistant professor of electrical engineering.