Hi Jan,
What is your real use case? A typical XML based workflow involves XSL-FO -> PDF route using FO processor (Antenna House, XEP, FOP etc).
I'm looking to use DITA to publish a literary journal. The journal will contain contributions from multiple authors. I also want to publish works by single authors. I won't make enough (or any) money to justify the very expensive license for Antenna House or XEP, thus they are not really an option. I find FOP to be really sub par. My basic work flow would have to be DITA XML > DITA OpenToolkit (build process) > ??? > ConTeXt > PDF, where ??? could be some Toolkit XML output, XHTML, HTML5, FO.
Anyway, there are several ways. If you are not locked to DITA yet, I would strongly recommend switching to DocBook instead :-)
The DocBook vocabulary does not fit my needs. TEI is much closer to what I want, but I'm already familiar with DITA.
This XSLT way is most natural for XML processing, but I understand that writing XSLT transformation is discouraging for many people.
Yes, I'm generally OK with working with XSLT. DITA has a build system, called the DITA OpenToolkit that is build with Ant, a bit of java, and XSLT. You can see the toolkit here: http://dita-ot.github.io/ The Toolkit already supports PDF output via FOP, but as I said, FOP leaves quite a bit to be desired. Best, Mica