Hi, I am working with a University research project that deals with automated verification of semi-structured data. The goal is to check for consistency criteria in documents that have a difficult structure, that were created by more than one person, or that were created over a long period of time. Possible criteria are e.g. "no topic is defined twice", "for every definition, there must be an example somewhere", or "every section, except for the introduction, must begin with a motivation and end with a summary" (those were formulated for e-learning documents that we worked with). These criteria are specified in a mixture of description logics and temporal logics, and validated using model checking against the document(s). Further information is available online at http://www.im.uni-passau.de/db/projekte?project=VerDiKt&lang=en (currently only in German, but some of the papers linked at the bottom of the page are in English). The scope of the project is two people working full-time, several people part-time, for 2-3 years. We are now looking for documents that can be used to test our theories and tools. Would it be possible for us to use the KDE docbook documents for this purpose (purely scientific)? I've had a look at svn://anonsvn.kde.org/home/kde/trunk/KDE/kdebase/doc, and the documents seem complex, well-written, well-maintained, and (in total) quite large. So they could be very useful to us. As a bonus for you, we might actually catch a few inconsistencies that the human eye has missed so far.
Kind regards, Chris
