W. Randolph Franklin rejoined RPI for the Fall 2002 semester after a rotation to the National Science Foundation in January 2000. At RPI, he is an Associate Professor in Electrical, Computer, and Systems Engineering, with courtesy joint appointments in Computer Science, and Decision Sciences and Engineering Systems. He is comanager with (G Nagy) of the Computational Geometry Lab, affiliated with the Center for Image Processing Research.
At NSF, he was Director of the Numeric, Symbolic, and Geometric Computation Program (later renamed to Graphics, Symbolic, and Geometric Computing (GSG), in the Directorate for Computer and Information Sciences and Engineering (CISE), Computer-Communications Research Division (C-CR). (In Oct 2003, C-CR disappeared in the CISE reorganization.) GSG supported research in computational geometry, computer graphics, numeric computing, mathematical optimization, symbolic computing, and automated theorem proving. It had 150 active awards. Frankin had additional responsibility for some panels and proposals in the Information Technology Research (ITR) and Integrative Graduate Education and Research Training (IGERT) programs.
Franklin was one of the prime movers of the two Computational Algorithms and Representations for Geometric Objects (CARGO) (FY02, FY03) solicitations, which were joint with NSF/MPS/DMS and DARPA/DSO, and wrote part of the DARPA Defense Science Office's GeoSpatial Terrain Analysis and Representations (Geo*) Future Area of Interest.
Franklin has held visiting positions in EECS at UC Berkeley, the US Army Topographic Engineering Center, Ft Belvoir, the Dipartimento di Informatica e Scienze dell'Informazione, Università degli Studi di Genova, Italy, the Dept. de Science Géodésique, University of Laval, Quebec City, Canada, the Division of Information Technology, Commonwealth Scientific and Industrial Research Organization, Canberra, Australia, and the Institute of Systems Science, National University of Singapore. He also helped found 2 defunct hi-tech startups, Hudson Data Systems, and Attic Graphics, Inc. He is an incorporator and board member of the Institute for Infrastructure Asset Management.
Franklin's degrees are from Toronto (BSc, Computer Science), and Harvard (AM & PhD, Mathematica Accomodata), where his advisor was Harry Lewis. His non-degree education includes Georgetown University's Governmental Affairs Institute's Congressional Operations Seminar .
Franklin's awards include an NSF Presidential Young Investigator award, an RPI Early Career Award, and being a significant contributor to the National Electrical Engineering Department Heads Association (NEEDHA) Innovative Program Award to RPI's ECSE Dept. for undergrad computer engineering studio course development. His professional service includes several site visits on behalf of the Computer Science Accreditation Board, membership in the Union College Computer Engineering Program advisory committee, and the usual reviewing and panelizing. He has developed and/or taught everything from freshman computer engineering, sophomore logic design, and senior computer graphics, to graduate software engineering and computational geometry. The graphics course was, at times, altho not a required course, the largest senior course in ECSE. He has been putting his course material on the Web since 1994.
Franklin's current research interests include computational cartography, computer graphics, computational geometry, and geographic information science, emphasizing small, simple, and fast data structures and algorithms, security and privacy, and computer engineering education.
His current goal is that anything he implements should work for N=1,000,000,000. His 3D connected component program does a universe of over 1000x1000x1000 voxels easily. His program to find mass properties of the union of lots of polygons (currently, squares) does 400M edges and will soon do 1G. This program demonstrates the functioning of several useful lower level operations, like edge intersection and point location.
A big reason for testing algorithms on large examples is that it's fun. However, it's possible that people aren't generally running large examples now since most current implementations are much too slow, when they don't crash. Large examples are an excellent stress test of the whole system. Also, this stuff scales down; when large examples take minutes, small examples run too fast to measure.
Franklin still gets occasional questions about PNPOLY, a subroutine he wrote in 1970 to test whether a polygon contains a point. Translated to C, PNPOLY has only eight lines of executable code. He has graduated 59 masters students, and ten PhD students, six of whom are now on the faculties of other universities.