...from the Microcomputer Power website.

The purpose of the Cornell Ecology Programs (CEP) series is to make public a number of computer programs used in our ecology research, dealing mainly with ordination and classification. The following list of programs is updated as new programs are completed. It is hoped that this series will save other ecologists the labor and expense of developing programs for these computations, many being rather routine. All programs are in IBM^TM's FORTRAN-IV. A subset of FORTRAN-IV has been used, intending to minimize conversion problems with other computers, and experience has indicated that most computer installations have minimal difficulties.

The CEP programs operate as regular, individual FORTRAN programs on mainframe computers. There is no supervisor program, no complex operating system, and no complex disk or tape usage. The programs place their data in high speed memory rather than on disk. Consequently, a mainframe computer is required to analyze large data sets having thousands of species and samples, but a small computer with 256K to 512K bytes of memory can handle small to medium data sets having a few hundred species and samples. Most microcomputer compilers implement FORTRAN-77 or a modification of FORTRAN-77, so some work must be expected to convert these programs for use with microcomputers.

[Newer development: see our CEP-PC MS-DOS^TM microcomputer package which is ready to run and includes FORTRAN-77 source.-ed]

An exceptional feature of several recent programs in the CEP series is that the computing workload rises approximately linearly with the amount of data. Most multivariate algorithms and programs have workloads rising as the square, cube, or higher power of the amount of data. Because of their linear workload, these CEP programs are readily applied to enormous data sets having thousands of species and samples. Frequently the computing time is on the order of the time required merely to read the data, or within a small constant times the reading time. The robustness and quality of the results have not been compromised in the effort to offer improved computing speed. These programs have been tested in relation to numerous other programs on a broad test data base with scores of simulated and field data sets. As shown in the articles cited under each program, the robustness and quality of the results are exceptional.

Individual topics such as non-hierarchical classification, hierarchical classification, and ordination are treated in the various articles cited here as well as in the program documentation. The general topic of multivariate analysis, however, is treated in an integrated manner in Multivariate Analysis in Community Ecology, by Hugh G. Gauch, Jr. (Cambridge University Press, 1982). Users of the CEP programs who are new to multivariate analysis may find this text a useful starting point.

The format for entering data matrices is explained most fully in the manual for ORDIFLEX. Numerous convenient formats are allowed. More recent programs, CEP-30 to CEP-41, use a single standardized format, but a utility program, CONDENSE, can copy any ORDIFLEX format into the standard format. The older programs, CEP-1 to CEP-21, use a slightly different format than ORDIFLEX, involving only a few trivial changes in a few lines.

Order Form/Price Sheet
This page last rendered at Ithaca, NY Thursday, August 14, 1997 2:52 AM