Article I · The promise
One ideal per category. Or none.
Our authority does not come from popularity, volume, an opaque score, or commercial influence. It comes from applying the same public rules to every case, showing the evidence, admitting uncertainty, and correcting the record without quietly erasing the past.
One product is named when it earns the standard. No ties or preference matrices that send the decision back to you.
The point is to end a reasonable search, not create another comparison spiral. A pick must be defensible, not mythical.
Expected failure must be repairable, rebuildable, maintainable, or rationally renewable.
We clarify the job, separate different uses, and test repair and renewal before saying that no product qualifies.
No ruling disappears silently. Confirmations, changes, and removals remain part of the record.
Article III · The questions
Five questions every product must answer.
An unanswered question is not a pass. Every final pick must clear all five questions, regardless of popularity, price, or review scores.
- G-01Is the category clear?
We must answer this question in the case file and support the answer with evidence.
- G-02Does the useful part hold up?
We must answer this question in the case file and support the answer with evidence.
- G-03Can failure be repaired or handled rationally?
We must answer this question in the case file and support the answer with evidence.
- G-04Will this model remain identifiable and serviceable?
We must answer this question in the case file and support the answer with evidence.
- G-05Does it give you one answer worth acting on?
We must answer this question in the case file and support the answer with evidence.
Article IV · Evidence standard
Show us how it lasts.
Three pages repeating one press release are one source, not three. We keep facts, our inferences, and unanswered questions distinct.
Article VI · Review status
Final picks are earned.
A public conclusion that has crossed the evidence floor.
Useful research statuses that are not recommendations.
Evidence strength describes the completeness and resilience of a case, not product quality or enthusiasm. No final decision may be issued below 4/5.
Article IX · Disagreement
If you disagree, show us the claim.
A useful challenge names the specific question, product claim, source, or change in the model that should be reconsidered. It states an objection someone else can check and supplies verifiable evidence.
This is not a voting system. Popularity, brand preference, a competing nomination, disagreement without evidence, or payment does not change a verdict.
Articles XI–XII · Boundaries
Independent by design. Accountable in public.
The rules and complete public decision remain free. A future paid guide may sell convenience and depth, never a different verdict or hidden decisive reasoning. The private Oracle may draft and critique; it cannot create evidence or publish a decision on its own.
Technical record
constitution/CONSTITUTION.mdcfbd7ae3a56ce31e…SHA256PI-CONSTITUTION