GFDL text corpus
The GFDL text corpus is simply the text corpus that is under GFDL. The GFDL itself ensures that the original sources of such texts cannot control the disposition of the material (this is called the right to fork). Accordingly there are competing text bases in various places, such as the Internet Encyclopedia (which we might strongly support as Wikipedia has joined the enemy projects, but is run with the assistance of Wikipedia developers, and therefore suspect in the eyes of trolls) and the Consumerium R&D wiki - where moral purchasing and comprehensive outcome and transparency are going to be more intensely examined than at any other GFDL-based service.
Trolls seek to achieve independence of user interface and access the GFDL text corpus itself directly through any user interface. This they are doing at present through use of tools they do not share with non-trolls, so little can be said about these. If this fails they may use legal means against software imperialism, e.g. GFDL clauses violated by Wikipedia, libel laws violated by sysop power structure which have turned Wikipedia into a libel pit.
One problem is that most GFDL text is at present in a data jail inside MySQL dumps, which is not even close to satisfying the GFDL clauses requiring source text access. While a decompression utility provides access to the strings, only one software suite (mediawiki) can actually present it, as there is no wikitext standard - a form of developer usurpation?
Eventually, by a combination of legal and technical means, trolls will succeed, and the GFDL text corpus will be liberated and visible and editable as a whole, and it will be relatively easy to undo sysop vandalism and avoid cabals.
At which point, a group something like m:The ideal Wikipedia board must be recruited to decide how best to evolve it to do some good for this planet...