In WinForms version the encoding is selected automatically by the MSXML API based on text content. I think we stick to UTF-16 in Java because our developers could not find a way to implement such auto-detection there. The encoding should be UTF-16 and not UCS-2 actually, we setup the following options in code for the stream writer and XML document:

transformer.setOutputProperty(OutputKeys.ENCODING, "UTF-16");
...
new OutputStreamWriter(stream, "UTF-16")

UTF-16 and UCS-2 are the same except for high-order codes (larger than 65536) according to http://www.coderanch.com/t/278588//java/writing-UCS-encoded-files-Java, so perhaps it's detected as UCS-2 by notepad++ because there are no such symbols in the file.

I hope that helps,
Stoyan

I don't think one or the other is more standard, and both .NET and Java store strings in UTF-16 format internally, so you can avoid some encoding conversions if you store the XML files in Unicode 16 too. You could use the saveToXml(Document) method to get the XML DOM tree in a Document object, and then save it in UTF-8 format if you prefer.

I hope that helps,
Stoyan