Hi, I found a solution, that is acceptable for me. Criticism and help welcome! It's mainly:
$ xsltproc --nonet --xinclude txt.xsl document.dbk | \ w3m -cols 70 -dump -T text/html -no-graph > document.txt (txt.xsl attached) - uses only free software - handles UTF-8, incl. CJK and RTL - handles tables - works with book, article, and refentry - runs on Linux (presumably under any OS) - runs non-interactively at the command line - features an ugly hack to get left margins for paragraphs - has a centered document title Thanks to everybody for hints!
txt.xsl
Description: XML document
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
