Problem:
I have an excel input file, generated on a German PC, i.e. the file was written 
with an ISO-8859-1 encoding. The file properties as well as the content (sheet 
name and cell content) contain German umlaute.

Then I'm reading this file with a java engine with -Dfile.encoding=ISO646-US. 
I'm doing this in a debugger (CodeGuide). When reading the document's 
SummaryInformation with HPFS the returned strings (Java Unicode) contain "?" 
instead of the umlaute. 
When reading the sheet name and cell content I see umlaute as expected.

No examplary output:
Unfortunately, I can see this only in the debugger. I do not know, how to show 
this with a short example. If I use the property -Dfile.encoding=ISO-8859-1 
then I get the correct result with umlaute. If I use another encoding (in my 
case ISO646-US), then a System.out.print() converts all Umlaute into "?".

System environment:
I have downloaded poi-bin-2.5-final-20040302.zip from http://ftp.uni-
erlangen.de/pub/mirrors/apache/jakarta/poi/release/bin. 
So I expect this to be version 2.5 (not in the list above).
I'm compiling and running everything with jdk 1.4.2_02.

Relevance:
Problem not only occurs with explict setting of file.encoding property but also 
if file will be read on a maschine with a different default encoding. We are 
only interested in the Java Unicode String, not in any other output device.

Further info:
The current HPFS sources in CVS contain a class "VariantSupport.java" which 
seems to implement codepage support in the SummaryInformation. This source is 
not contained in the downloaded 2.5 version.
I can provide an example if needed, I have no idea how to attach it here.

Best regards,

   Michael Gesmann