I've been horribly busy and have a backlog of things I want to blog about, but I've just had a letter from someone wanting to know about how the talk in Cottbus went, so here goes.
The BTU Cottbus is the university that granted a doctorate to a Vattenfall manager, Detlev Dähnert, on the basis of a dissertation that has large blocks of text parallels on 44 % of the pages, as has been documented by VroniPlag Wiki, and for which at least some of the research was done by the company Infratest, as a former worker for the company has stated (third portion).
The university deliberated and announced that they were not rescinding the doctorate, as there were only "technical weaknesses" in the thesis. I, and others, have been calling on the university to publish the expertise about the thesis. One would suppose that this expertise would list the 125 fragments of text parallel that VroniPlag Wiki documents and explain for each one what exactly the weakness is and why this is not to be considered plagiarism. I, for one, would fail a student on pretty much just one such fragment.
One fresh graduate of the university, who was not planning a career in academia, gave a talk in July in which he concluded that there was extensive plagiarism in the dissertation and called on the university to re-think its position. He is now no longer with the university (his contract ran out, as is normal for graduate student contracts in Germany), but a group of students contacted me and asked that I continue the discussion in the winter term.
There appeared to have been a bit of a scuffle between the university and the students on permission to use a room, but I was allowed to speak. There were only 18 persons there, including one member of the commission. I spoke on many cases of plagiarism in dissertations, including many historic cases and some VroniPlag Wiki cases (slides, 16 MB, the slide numbers 32-35 include excerpts from Dähnerts dissertation).
The discussion was excellent -- there were many issues raised on how to go about determining plagiarism, why the need to demonstrate intent should be necessary, remarks that the intent is clear if you actually look at the text parallels, comments that one of the persons plagiarized didn't feel cheated so it must be alright, and a general criticism of the numerical fixation at VroniPlag Wiki and the barcode representation that overly abridges the issues involved.