The PreProc is an input filter that changes the Body Area of the source document. It is a "find and replace" feature, applied right after the line is read from the document source, before any parsing by txt2tags.
It is useful to define some abbreviations for common typed text, as:
%!preproc: JJS "John J. Smith" %!preproc: RELEASE_DATE "2003-05-01" %!preproc: BULLET "[images/tiny/bullet_blue.png]"
So the user can write a line like:
Hi, I'm JJS. Today is RELEASE_DATE.
And txt2tags will "see" this line as:
Hi, I'm John J. Smith. Today is 2003-05-01.
This filter is a component that acts between the document author and the txt2tags conversion. It's like a first conversion before the "real" one. This behavior is similar to an external Sed/Perl filter, called this way:
$ cat file.t2t | preproc-script.sh | txt2tags -
So the txt2tags parsing will begin after all the PreProc substitutions were applied.
Note: Remember that the preprocessing is applied only to the BODY of the source document, not including the Header Area and Config Area.