I'm still not clear on a few key points -
1) Is your output going to be pure text, or is it going to be xml with the logic formulae contained in the xml elements?
2) What application or code will consume your result in the next stage of processing?
The answer to 1) is crucial, for this reason - you MUST escape the "<" characters in character content (though you can surround a formula in a CDATA section instead of escaping the character individually). On the other hand, of you are creating pure text, no escapes are needed.
If you are creating xml output, the processor in the next stage should unescape all characters as needed, but if you are extracting the content using, say regular expressions then of course that wouldn't happen.
If you want to create plain text output, just set the method attribute in xsl
utput to "text" in your xslt stylesheet, and the results won;t be escaped.
As for the "PI_DISABLE_OUTPUT_ESCAPING" constant, I have never used it but from the javadocs, it looks like the processor just inserts a processing instruction, which might be understood by some downstream code but not by most. That's probably a red herring in your case.
You said "the problem is that at WRITING TIME the parser escapes all the ">" characters". Actually, the
parser does not create the output. The transformer is doing the output. The parser is the code that reads the xml document and analyzes it to find its structure.
[BTW, Harold, you sent me a private message but your profile does not allow you to receive private messages so I could not reply except here]