Bowser wrote:
...I love the term "unconstrained variables". Seems to fit a lot of circumstances which may have shorter, less polite, descriptions.

Alex writes:
Actually, it's just a fancy way for scientists to say that models have numeric values with large error bars. For example, the measurement of Mars' precession rate (and by extension, its moment of inertia) *constrains* the size of its core, depending on the accuracy of the measurement. Mars Pathfinder gave a good measurement that helped bring this variable down to bite size, but more accuracy (e.g., from 2007 NetLander) is needed to bring it down further to help distinguish various Mars interior models.

Alex R. Blackwell
University of Hawaii