Start Here

Five files, in order

You do not need the whole archive to find out whether this is worth your time. You need five files, and you need to know what each one is for.

Each of these demonstrates a different rule we hold ourselves to. Two of them are rules about our own mistakes. Read in order they answer the question that actually matters first, which is not what is this about but can I trust it.

  1. A real photograph proves the photograph is real. It proves nothing about the caption.

    NASA photographed the “Black Knight” satellite — and catalogues it as space debris

    FalseSF-0001

    Start here because it separates the two things people fuse together. The image is genuine, unretouched, and in NASA’s public archive. The story attached to it is false, and the refutation is not an argument — it is a field in NASA’s own catalogue record for that exact frame.

  2. When the answer is not settled, the file says so instead of picking the popular one.

    Nobody has settled how Devils Tower formed, and the Park Service says so

    MisleadingSF-0022

    Second, because a publication that only ever resolves things is one you cannot trust when it resolves something. Here the Park Service lists four explanations and writes that discrepancies remain for all of them — so this file reports an open question, and the verdict falls on the confidence rather than the claim.

  3. The correction overshoots too, and we guard against our own.

    Grand Prismatic's rings are microbial. Its blue centre is physics, not sterility

    Mostly TrueSF-0016

    Third, because this is the failure mode debunking never checks for. The rings really are microbial. The blue centre really is optics rather than pigment. And the tidy conclusion that follows — that the middle must therefore be dead — is also wrong, because the Park Service records abundant Archaea living in it.

  4. The absurd file is sourced exactly as hard as the serious one.

    The comb-over patent folds three sections of hair over a person's hand

    VerifiedSF-0013

    Fourth, because by now it would be fair to assume this place is joyless. It is not. This is the patented comb-over, and the reason to read it is that the claims are stranger than the joke it travels as: claim 3 requires an object held over the hair mid-fold, and claim 4 specifies that the object is a person’s hand.

  5. A single cause is usually a sequence, and naming the sequence is the work.

    Lead began Picher's relocation. Collapse risk widened it. A tornado hastened it

    Mostly TrueSF-0017

    Last, because it is the hardest discipline in the archive. The famous version of Picher has one cause. The record has three, arriving in order — and the file refuses both to rank them and to add a fourth that sounds plausible but is not in any source we read.

After that

If the five did their job you now know the rules, and the rest of the archive is just more of the work. The floor plan shows every subject and every verdict with honest counts. The collections are the files read as arguments rather than as a list. Or take one at random.

If instead you want to know how any of this is decided before you read another word, that is the method — including the gates a file has to pass to be published at all.