One challenge in DITA is that now translators have to sort through a huge number of files to translate all the content. Right now most software sorts them alphabetically or by some order that isn't logical in the context of a publication. It would be easier and produce higher quality translations if all the content (only objects that weren't already translated, obviously) for a particular chapter were placed into a single XML file to be translated. I saw something similar a while ago: https://sourceforge.net/projects/ditaxliff/files/ [presented by Bryan Schnabel at an SDL conference a couple of years ago]