[tlug] iconv / Python / unicode question

[Frank Bennett <bennett@example.com>]

I am working on a Web design based on python tools (Zope CMF),
and it needs to be able to convert documents from Japanese Word
on the fly, for display.

The conversion is failing on my first test document, because it
contains characters that both iconv and the Python unicode object and
Japanese codecs think are invalid.

I am happy to kill the characters, but I need to get the rest
of the content onto the Web.

Is there a toggle in the python unicode object that will just
drop non-conforming characters on the floor?  Failing that, is
there __any__ filter that will strip out these blocking characters
from a file, so that it can be run through these tools without
blowing them up?

Hope someone can help,
Frank