Quote:
Originally Posted by llasram
Hardly scientific, but I downloaded the torrents for a handful of fairly large pirated e-book libraries. All the broken DRM e-book formats are HTML-based, so I extracted just the HTML books (had ‘htm|HTM’ in the filename). From that set I took a random sample of 100 which I individually examined to see if there was any evidence they had been derived from pirating the e-book edition of the work. And the number derived from pirated e-books...
None. Every single one was obviously made via scanning and OCRisg.
|
Slightly
but it is quite an interesting result. I think it is scientific enough. It sounds like a valid statistical test judging from the 'random sampling'.
.