re: #220 Obdicut

I’m sorry, but doing forensics on code based on what was included in some folders in some document dump is insane.

Not so much, really. In fact, part of my job includes what we refer to as “code spelunking,” diving into year-old, undocumented code whose author has long since vanished, trying to figure out what it does and how it does it. It isn’t exactly pleasant work, but it’s occasionally necessary and always well within the realm of the possible.

I haven’t looked at the code in question - because I spend enough time doing such things - but it ought to be possible to figure out how it works and whether it accurately calculates results based on either published models or the models documented in the rest of the material. It would also be almost trivial to determine just what sort, if any, data filtering is done on the input data before it reaches the model calculations themselves.

I haven’t heard that anyone’s taken such a step yet.