Major Software Developments

  1. CROSSWALK (1985-87). A simulation of extramolecular configuration in crystalline polymers, called state-of-the-art by DOE reviewers. Instrumental in 5 publications and 6 invited lectures. Supported by the US Office of Naval Research.

  2. PolyStruct (1987-90). A PC-based suite for calculating various physical and topological structural variables for prediction of material properties of crystalline polymers. Supported by the US Office of Naval Research.

  3. KNODES (1988-89). A suite of workstation software for use by scientists and engineers working on reactive flow problems numerically. The modules replace human assistants for many tasks related to the use of large mainframes (supercomputers) for numerically intensive investigations. Supported by the Florida High Technology and Industry Council (design/prototype.)

  4. ENBP (1991-92). Expert Network BackPropagation learning with influence factors. Portable C version developed with S. I. Hruska. CM FORTRAN version developed with S. I. Hruska and K.D. Nguyen. Supported by the Florida High Technology and Industry Council.

  5. xnet (2003-10). Modernized versions of ENBP developed in C++, using adjacency list representations for networks. Improved runtime of the learning algorithm to O(n) per iteration for typical sparse expert networks.
    Inter-operable with companion UI/Controler written in Java, including rule editor, reasoner, explanation module, and learning module, developed by Rich Waddington of SUTI, Inc.

  6. fsuSTL (1998-2013). A standard template library meeting most of the specifications and all of the functionality requirements of the C++ STL, but with some naming and conventional changes making the library sufficiently different from the STL to be suitable for student assignments. Students can build, expand, experiment with, and create applications using the library. The library also facilitates exploration of issues, options, and choices that were faced by the STL standards committee. Suitable for use in modern data structures and algorithms classes and also in professional development environments.

    The library features improved set, map, and hashtable containers that are adaptor classes, allowing the client program to select the underlying implementation technology while maintaining the abstract data type through the common adaptor interface.


Back to Lacher home page