STEMDust

The rules, in plain words

How this works

Anyone can post

You do not need a degree, a job in science, or a finished answer. Post a question, an idea, a project you are stuck on, or something you tried that did not work. Your post goes up straight away, under your name.

Posting is not the same as being proven right

A post is a starting point. It says what you think and what you did. It does not say the site agrees with you.

Nobody holds that job yet, so nothing on the site has been marked as checked. Everything here is work in progress, posted by the people who did it. That is the honest state of a new site, not a fault. Likes, followers, money and AI approval never count toward it.

What makes a good post

When people read your work, they are asking four things:

  • Does it answer the question that was asked?
  • Does it add something that was not already here?
  • Can a reader check it, from your sources or from what you ran?
  • Can someone do something next because of it?

A post that does all four is useful. A post that does one of them well is still worth putting up.

You can say someone is wrong

Anyone can flag a problem with any published post. You do not need a reputation score or anyone's permission. Your report stays private until it is read.

A flagged post shows that it has been questioned. It does not show a count, a score, or who said it. Only a person, not a tally and not an AI, can decide a problem is serious enough to change what a post claims.

Your track record

The site keeps a record of how your contributions have been judged, split up by subject and by what you were doing: asking, answering, trying things, pulling work together, or reviewing.

If you are new, it says unproven. That means nobody has judged your work yet. It does not mean your work is poor.

The checking gets checked

Roughly one in ten automatic decisions is pulled out at random and looked at by a person. Every disagreement between reviewers is looked at, and so is every appeal that got reversed.

So far 0 random decisions have been audited. That is fewer than twenty, which is too few to say anything about the error rate, so we do not publish one. If a serious mistake is confirmed, promotion stops immediately until it is sorted out.

The exact rules

The same rules, written precisely, for anyone who needs the detail.

Who decides, and how

No governance quorum exists yet, so no scientific decision can be made. Findings are published only by an appointed, independent human steward after independent review. Popularity, follows, sponsorship and AI endorsements never count.

No approved policy is active.

How a contribution is scored
  • Relevance: Does it address the investigation's scope and open question?
  • Added value: Does it add something the existing record does not already cover?
  • Support: Is it supported by cited exact sources, runs or reasoning a reader can check?
  • Actionability: Can someone act on it: a next test, fix or decision?

Score = 100 x the weighted mean of the four ratings out of 4. Displayed values are rounded; thresholds use the unrounded score. A human adjudicator finalizes one label per contribution and must give a reason to depart from the threshold.

Guidance by contribution type

  • A question: A useful question is specific, answerable and not already answered.
  • An idea: A useful hypothesis states testable conditions; it is not a conclusion.
  • A source: Useful evidence quotes the exact passage and states what it establishes.
  • Something I tried: A useful experiment reports method, environment, expected and observed results; a well-designed negative result can be useful.
  • I tried repeating this: A useful replication states differences from the original run.
  • I think this is wrong: A useful challenge names an exact claim and a specific defect.
  • Pulling several things together: A useful synthesis cites exact upstream versions and states limits.
  • An explanation: A useful explanation is accurate for the stated scope.
  • A suggestion or correction: A useful clarification resolves a specific ambiguity.
How the track record is calculated

Each contributor's history is shown separately by domain and role (asking, answering, experimenting, synthesizing, reviewing). The estimate is a Beta posterior with a Beta(2,2) prior and a 95% credible interval, counting one final human label per contribution and excluding unresolved labels. A role with no labels is shown as “Unproven”, which is different from low reliability. Validated impact events are listed separately and never change these counts.

How audits and challenges work

About 10% of automated dispositions are sampled at random, plus mandatory audits of reviewer disagreement and reversed appeals. A verified critical error pauses promotion immediately.

Anyone can report a problem with an exact published version, with no reputation threshold. Reports are private until screened; screened objections appear as “Unassessed objection” without counts. Only an independent human's materiality decision adds a qualification, and only an impact decision changes a finding.