Wow, that's about the longest "readme"file I've ever encountered. Not what I expected, in that a formal readme file is documentation, and should list known bugs/issues, and provide handling/usage instructions e.g. installation. These are raw notes, which are more valuable in a forensic investigation like this.

I don't grasp the fine details, which are written in project-level jargon; nor the next layer of stats-geek jargon. I can pick up a general sense of what's going on at the outermost general programming layer.

I may never finish it, but the first part appears to describe a project to recover raw data from published data, which, I believe, was one facet of this controversy. It's possible to recover the raw data if the processing performed was some kind of correction against known constants, for example, instrument calibration data, interpolation between units of measure or instrument technologies (mercury vs. digital themometer e.g.), etc etc. The "suspicious" code could be removing an array of correction factors that were originally applied to the raw data. (I'm speculating here.)

After that, the poor bastard appears to spend weeks tracking down lost files on whatever computer system he had the misfortune to inherit. That's still where I am, and, chillingly, the scroll bar says I"m only 10% in. Eeeeeeeee....I'll see how much more I can browse later.