readCEP can return data in "long" format as triplets of row names
(typically sites), column names (typically species) and data
values. As default, rows with zero or negative values are omitted.
Reports used Fortran compiler (e.g., in 00install.out) per wish
of issue #8.
Can return the data as a sparse matrix (of Matrix package). This can offer considerable saving in disk storage. However, several R functions may be unable to handle sparse matrices, but they must be changed to ordinary dense matrices or data frames. Moreover, sparse matrices are often changed to full dense matrices within the analysis. For instance, a centred or standardized data matrix in ordination is dense.