In a project with a German government authority that
runs since 1998 we use spam to normalize SGML files
(because a lot of information is in the DTDs, in particular
default values for attributes).
There is quite often the need to process more than one
SGML file, ranging from a few dozen to several
thousands. It is possible to process them one by one, of
course; we do this for quite some time now. But it is a
major performance overhead to call spam (via "system"
from our Perl program) for every file. If there were a
possibility to give spam arguments for an input and an
output directory and it would simply process every file in
the input directory and write a result file with the same
name to the output directory - this would speed up our
publishing process very significantly.
Being not very fluent in C++ and having no Visual C++
IDE, I'm hesitant to try this myself. But it doesn't sound
too difficult for the Jade/SP wizards.
Log in to post a comment.