Le 13/06/2014 11:57, pygmee a ?crit : > I guess the result will always be better if done by the software itself than > by a conversion tool. For example how will calibre deal with page numbers, > footnotes, image captions... there many cases where a conversion tool can not > understand the layout, especially when it is done from PDF which is not > semantic. Sure. How can the export recognize that some specific text is the legend of that other specific image frame ? Just because it's closely placed in the PDF ? Doesnt'it requires adding some semantic data, for example defining the attributes that scribus enables to define for frames ?
JLuc
