Within IMG, MATLAB codes are used extensively to predict device performance, usually via lumped element models (LEMs). As these models increase in complexity, it becomes harder to keep track of the many workspace variables, function input/output syntax, data formats, etc. It can also be extremely difficult to save simulation results or post-process in organized ways.
With object-oriented programming (OOP) in MATLAB (updated significantly in R2008a), complex LEMs can be encapsulated in a series of objects (with their own properties and methods) that free the end-user from many of these difficulties. The seminar, given by Matt Williams, will include a brief introduction to OOP, how it can be naturally applied to LEMs (via a simple example), and a short demo of a more complex implementation (my own LEM of a piezoelectric MEMS mic).