This should be nothing new to you when I say "Don't trust your eyes".
But specifically when it comes to Unicode, I feel like saying it again: "Really don't".

...at least for Python 3.0.

Still hoping we can only wonder about sentences like the following in 10 years:
Some languages use special characters (Chinese, Japanese, Arabic, Klingon, etc.) that are difficult to handle with traditional software.

[img_assist|nid=138|title=I�Unicode|link=node|align=left|width=460|height=83]

Stole this somewhere, I'll provide a link once I find it again.
See Unicode Specials on english Wikipedia for the "replacement character".

Python (2.4 and 2.5 tested) seems to have problems when an exception is thrown that contains non-ASCII text as a message.

In my case I constructed an exception like

>>> try:
...     raise Exception(u'Error when printing ü')
... except Exception, e:
...     print e
...
Traceback (most recent call last):
  File "", line 4, in ?

If you are at the beginning of a new project and are currently considering which programming language to use, and furthermore you will definitely need Unicode support, then please consider some other languages than python.

Don't use python. Really only used it, if you have no other choice.