
    File MODCODE is no longer a schematic program for putting externally
computed correlations into a file that MODA accepts as input, but an
ordinary-language description of the form this needs to take.  If your
correlations -- or any rescaling of them by a global constant to omit
decimal points -- are in a textfile, you can easily use a text-editor
to shape that into MODA-readable form. >>> NOTE: You must name this
file with extension COV or MODA will not recognize its existence.

MODA accepts any ASCII file as input if its name-extension is COV and
its content comprises two, three, or four blocks of consecutive lines
that can include or be separated by blank lines:

1st block (optional): One or more header lines describing this file's
    contents. No header line may begin with a numeral, but otherwise
    they can say anything you want.

2nd block: The first and only line of this must start with the number
    NV of variables whose standardized covariances are to be entered,
    optionally followed by the numerical value of NV*(NV+1)/2 which
    MODA will compute if you don't supply it.  COV-files from HYDATA
    have additional entries on this line, but MODA doesn't require those.

3rd block: This must list the NV*(NV+1)/2 entries in the lower triangle
    of your correlation matrix.  You can enter these in NV lines whose
    entries are of increasing length forming a triangular display, or you
    can concatenate those lines as one long string with line breaks at
    your convenience.  Entries in this string must of course be separated
    by spaces or commas.  And decimal points can be omitted regardless of
    how many places you have computed so long as all are rounded to the
    same decimal.

4th block (optional): This can contain one or more sub-blocks in any
    order, each starting with a text line whose first letter keys the
    content it describes, and followed by one or more lines listing this
    content.  The sub-block of main interest comprises names of the input
    variables:  The header line for these must start with "N" and be
    followed by one or more lines listing names for NV variables in order
    of their occurrence in the correlation matrix. Names should be separated
    just by spaces and must begin with a letter, contain no more than eight
    alphanumeric characters, and not contain a letter after any numeral.
    The same name can recur, such as " ??? " when you aren't sure.  If no
    names are entered, MODA will supply numeric defaults for them.  See
    any COV-file written by HYDATA for more details on this and other
    block-4 entries.

