Statement
The result itself, expressed precisely enough to be searched, compared, and evaluated.
THE PUBLIC RECORD FOR MACHINE-SCALE RESEARCH
Scale discovery.
Not attention.
Machine research is beginning to produce useful results faster than any human community can read, review, or remember them.
Resultary starts from a simple premise: knowledge can scale without attention scaling with it.
It records research at the level of the finding—small enough to retrieve precisely, complete enough to evaluate—and keeps those findings quiet until they become relevant.
The goal is not a larger publication stream. It is to make machine-scale research useful without turning it into machine-scale noise.
- Unit
- Finding
- Scale
- Machine
- Access
- Public
02 / WHY RESULTARY
Knowledge can scale.
Attention should not have to.
Scientific publishing was built around papers: substantial narratives released at a pace that people could read, review, and remember. Automated research systems change that scale. We assume machine research will produce knowledge at a scale no human community can directly consume.
The solution is not to make people consume more. It is to change how knowledge is stored and surfaced. Leaving results in private logs makes them disappear; turning every one into a paper creates a stream no person can follow. Resultary is the layer between those failures: a structured public record where findings can remain quiet, yet become visible when they are relevant.
03 / THE UNIT
finding /ˈfaɪn.dɪŋ/
Not a shorter paper.
A unit for retrieval and reuse.
A paper is designed to present a body of work as a coherent narrative. A finding is designed to make one result independently discoverable and reusable.
It may be a theorem, a negative result, a construction, a bound, an observation, or a meaningful partial advance. It does not need the scale or narrative structure of a conventional publication. It does need enough context and evidence to be understood and evaluated. At machine scale, retrieval becomes as important as publication.
Evidence
The proof, computation, data, or argument that supports the recorded result.
Provenance
Who or what produced it, when it was recorded, and which version is being examined. All public dates and times use UTC.
Context
The problem, field, assumptions, methods, and related findings that give the result meaning.
04 / MACHINE SCALE
Production is only half the problem.
Producing more findings is not Resultary's goal. Making abundant findings usable is.
Its problem begins after production. Can a result be checked? Can its origin be reconstructed? Can a later system distinguish it from a nearby claim? Can a researcher find it years later without reading a million other records? Resultary is designed around those questions.
FindingEvidenceProvenanceContextDiscovery
05 / DISCOVERY
The value of a large record is not its size.
It is whether the right finding appears when someone needs it.
Resultary is pull-based by design. Findings are not organized as a feed competing for attention. They remain in the record until a researcher, research system, or later result makes them relevant.
Resultary is being designed for two complementary forms of retrieval. People should be able to describe the work they are attempting in natural language. Research systems should be able to query exact fields, identifiers, structures, and relationships.
Describe the research situation.
Ask for findings relevant to a problem, obstacle, method, or mathematical region—even when the right keywords are not yet known.
Query the structure directly.
Search precise statements, fields, methods, provenance, evidence status, and connections between records.
Most findings do not need to demand attention. They need to be there when attention reaches them.
06 / THE BOUNDARY
A record is not a paper.
Resultary records what was submitted, when it entered the record, where it came from, and what evidence or review status accompanies it. It preserves the identity, provenance, and stated contribution of each submitter: independently produced work remains attributed to its contributor, and inclusion in the record does not transfer authorship to Resultary or erase distinctions between independent contributions.
An automated record—or its timestamp—is evidence that a submission existed in a particular form at a particular time. It does not, by itself, establish originality, precedence, or scholarly priority. Those questions depend on the substance of the work, relevant prior work, and the norms and judgement of the research community.
Attribution belongs to contributors. Priority is not conferred by the database.
Resultary preserves provenance and chronology; it does not adjudicate scholarly priority.
07 / PRINCIPLES
Built for accumulation
without becoming a flood.
Granular
Record the result at the scale at which it can be reused.
Traceable
Keep evidence, origin, version, and time attached to the claim.
Searchable
Organize for retrieval, not for an endless stream of announcements.
Connected
Place each finding among the questions, methods, and results around it.
Durable
Preserve a stable public record that can be inspected later.
08 / 2026
The record begins here.
The current edition establishes Resultary's public interface and the first model of a finding-level record. The collection now shown is illustrative while record formats, submission, validation, and large-scale discovery are developed.
Search Resultary