Monday, April 25, 2005, 5:34:04 AM, Andreas Jung wrote:
One last note from myside. I have experiences with unicode since over 7 years while working with multilingual documents in the e-publishing business. It is good practice to perform *any* unicode related work *only* on unicode datatypes (Python unicode strings!!!) and *not* some byte-encoded unicode strings as utf8 or whatever. These encodings should only be used on the output level when presenting unicode data to the user - either through-the-web, as export format etc. This is a strong advice you should follow.
As someone who works often with Java I absolutely agree with it. Just I don't know how to do it with Zope/Plone/other 3rd party products (not written by me), since they use not unicode strings. I don't know, maybe it can be specified for Python that it uses unicode for plain strings as well, but anybody has successfully done that with Zope? Like, what about the C parts of Zope then?
-aj
-- Best regards, Daniel Dekany