Work on what was to become “structured design” began in the early 1960s. Structured design, as a well‐defined and named concept, did not achieve appreciable visibility until the publication of an article in the IBM Systems Journal in 1974. There is more than one way to accomplish a “structured design.” Like structured design, the term object‐oriented design (OOD) means different things to different people. For example, OOD has been used to imply such things as The design of individual objects, and/or the design of the individual methods contained in those objects The design of an inheritance (specialization) hierarchy of objects The design of a library of reusable objects The process of specifying and coding of an entire object‐oriented application The term nonformal is used to describe approaches to OOD that are not well defined, step‐by‐step, or repeatable, such as those that emphasize the design of individual objects, specialization (inheritance) hierarchies, and libraries of objects. A survey of such approaches indicates that there may indeed be some repeatable rigor (and some sage advice) given for these approaches, but they are severely lacking when it comes to defining the software architecture of large and critical systems. Different approaches, techniques and experienced with object oriented design are discussed here.

No abstract

No abstract

We have carried out a careful study of He–H2 vibrationally inelastic collisions based on five different analytic fits to the Gordon-Secrest [J. Chem. Phys. 52, 120 (1970)] ab initio interaction potential. Within a model which freezes the atom-diatom orientation, we solve the collision dynamics fully quantum mechanically. Both for the 0→1 and 3→4 vibrational transitions, we find that the use of different analytic surfaces leads to substantial qualitative and quantitative differences in the calculated transition probabilities. In particular, we observe an anomalous angular dependence in the case of the original Gordon-Secrest analytic surface. We show that this arises from difficulties in the extrapolation of the ab initio points in the classically forbidden region at small center-of-mass separation. These results have important implications for future theoretical study of molecular collisions.

No abstract