I'm going to run into toolserver some simple python + djvuLivre routines to test the possibility to obtain a "wikicaptcha", built to be useful for wikisource activity.
I'm far from sufficiently skilled to write all the project, in particular the final user interface; but I'm not far to implement something like a "voluntary wikicaptcha", t.i. to select controversial OCR interpretation of words of a djvu file, and to present their image, extracted from image layer, into a html form, so that a willing user could upload their "human interpretation" and fix djvu text layer.
While asking if any of you is interested about, I wonder there would be any wrong, or hurting toolserver politics, or raising safety issues, in publishing the python code of such layman tries into toolserver wiki into a subpage of my account, or otherwhere, so that any toolserver user could take a look if curious or simply could take inspiration to develop the idea as the idea IMHO deserves.
Alex brollo