Documentation updated

lib/python/Products/ZCatalog/README.txt

@@ -36,5 +36,19 @@ ZCatalog
        The UnicodeSplitter will not be the default splitter so 
        the standard vocabulary of the ZCatalog will not support
        unicode except it is replaced with a new vocabulary using
-       the unicode splitter.
-
+       the unicode splitter. When using an unicode-enabled TextIndex
+       it is not neccessary that the corresponding attribute/method
+       of the indexed object must be a unicode string. If it is
+       a standard normal string then this string is converted
+       to unicode using ISO-8859-1 encoding. A different encoding
+       can be defined by an attribute '<attribute>_encoding' providing
+       the encoding.
+
+
+     - Using a KeywordIndex with unicode strings is slightly different.
+       In general you can mix columns containing unicode strings and
+       standard Python string to be indexed using a KeywordIndex. For
+       internal reasons it is neccessary to change Pythons default
+       encoding (set in site.py of the Python installation) if any
+       to the keyword uses a non-ascii encoding (e.g. using accented
+       characters).