The real discussion is much more nuanced and scientific than "peer review is bullshit," but that blunt approach is appropriate for Slashdot.

I totally agree. The problem is even worse in free software than in real science. It is enough for a "reputable" developer to state that a given free software project is "superior" to another for most everybody accepting it on his word alone, even if the evidence shows that the opposite is true.

Xz format inadequate for long-term archiving

Here is a better link with the originals and a link to katrillions of raw images.

It is not a better link; it requires javascript to even see it.

And that's your answer. A filesystem like FAT32 or ISOFS that's likely to still be implemented in future OSes and a recovery files which let you rebuild anything that suffers from bit rot.

Lzip and lziprecover can help you keep your data safe in the long term. (For the kind of data where it makes sense to use a lossless compressor, that is).

That does NOT eliminate the need for good code design, actually, having a good design phase is absolutely crucial to this approach, since else your juniors have to design. That would be ... let's say sub-optimal.

Even more important than good code design is good data format design. No amount of good code can compensate for the defects in a bad format.

I mean, why can't we keep unproductive people around? What's the limiting factor?

Global warming?

Where did you learn about that?

In the manual of any argument parser.

Without patents we would have trade secrets.

Then outlaw trade secrets by requiring that the schematic diagram, source code, recipe, etc, of anything sold to the public must be made public.

What, in your opinion, makes NTFS a pain in your ass?

For example performance with very large files.