I'm trying out the command line version of OpenConvert on the recently released British Library alto xml, aiming to turn it into plain text.
Although I have OpenConvert working on a sample (test4.xml, attached to this report in zip format), the text is very badly formatted. Each word is on a separate line, with 6 tabs before it, a blank line between each word, and 454 (!) blank lines between the header information and the first page text.
The online version of OpenConvert is formats the text very well.
(Using Mac OS X 10.11.6, & Java 1.8.112 in case that matters.)
test4.xml.zip