Mmhh why not distribute the load among various machines?
What does this process do? Parse SGMLs and translate them
them into HTML docs?
Let me look at the code (just tell me where it is, if it's public)
and I'll try to see if I can come up with some sort of distributed procesing script for it.
We have many servers here that could donate some CPU time
during the night and weekends. Perhaps this could render the docs
on a weekly basis. Given enough cpu-time donations perhaps the
docs could be generated on a daily basis.
The base of the code run to build a manual is in phpdoc CVS.
./configure --with-lang=LANGCODE_TO_BUILD
make mirror-files
Goba