![](https://secure.gravatar.com/avatar/5a39f1335b6eff433673ed987859fb24.jpg?s=120&d=mm&r=g)
Not only Hans van der Meer is producing (or trying to produce) ePubs at the moment; I reviewed my XSLT workflow (https://wiki.contextgarden.net/Epub_Sample) since my publisher colleagues and our distributor aren’t happy about just publishing PDFs as eBooks. While the first book in a series of four is now ready with a lot of manual corrections, in the second one the XML structure is messed up: the chapter title structure comes only after the chapter content, and index entry tags are empty with the actual entry behind like this: <paragraph>Ein sentimentales Rührstück ist <highlight detail="emph">Rico, Oskar und die Tieferschatten</highlight> dennoch nicht. Die beiden Hauptdarsteller <registerlocation detail="index" internal="396"></registerlocation>Anton Petzold (Rico) und <registerlocation detail="index" internal="397"></registerlocation>Juri Winkler (Oskar) füllen ihre Rollen ganz wunderbar aus...</paragraph> In ConTeXt code I marked the actors with \def\NPers#1#2{\index{#2, #1}#1 #2} as \NPers{Anton}{Petzold}. But since that just uses \index, the XML tag should catch the contents, don’t you think? I’ll come up with a MWE; don’t know if I’ll manage to reproduce the title/contents mixup though. The other book’s export is ok within the same project. It uses the same environment, but a slightly differing chapter setup. Any hints which setting could cause the re-ordering? (And I use \startchapter … \stopchapter of course.) Best, Hraban