Knowledge Universe › Medicine › Contribute
Contribute · conceptContribute knowledge — not just files.
A person can add real knowledge to the library even when a source file can never be shared. Before anything becomes community material we ask three questions — is it already known, what do the rights allow, is it good enough — and credit comes after those answers, never before.
International learning concept — currently available in English. DE / RU / UK: roadmap.
How contribution works
Five plain sentences.
- Bring a source or knowledge. A book, an open article, your own notes, a correction, a topic map.
- We first check whether it is already known. On your device, before anything is sent.
- We determine what the rights allow. Per source, per licence — recorded as a rights profile.
- Useful knowledge can still be contributed even if the file cannot be shared. Metadata, mapping, reviews, original explanations.
- Accepted community material stays linked to its source and rights status. Always.
Central principle
Contribution ≠ file upload.
Most valuable contributions are not files at all. All of these count, each with its own review and its own credit:
If a copyrighted book cannot enter the Community Library, the answer is never just "rejected". It is: "We cannot add this file to the Community Library, but your knowledge contribution can still help." — add metadata · review · classify · map topics · create original learning material.
When a file itself can be hosted
Accepted file bases (future).
Public domain
Copyright expired.
Compatible CC / Open Access
CC BY, CC BY-SA, openly published works.
Own work with granted rights
Notes, checklists, translations, corrections you authored.
Rightsholder permission
A confirmed right to share from the rightsholder.
Metadata / link
ISBN, title, edition and a link to a legal source — no file.
A normal purchased, copyrighted book is not automatically eligible for file hosting. Its metadata and a legal link are still a real contribution. What you may keep for yourself in a future My Study Space is a separate, rights- and jurisdiction-dependent question.
Contribution lifecycle · scripted demo
Pick an example and step through it.
Four prepared examples. Nothing is uploaded — every step is scripted to show how the real lifecycle would decide. Submitted ≠ public: a file never becomes Community Library content just because someone selected it.
- 1Discovered—
- 2Local duplicate pre-check (fingerprint in the browser)—
- 3Submitted — not public—
- 4Quarantined — held, invisible to the library—
- 5Duplicate check (3 levels)—
- 6Rights check → rights profile—
- 7Quality check (human review)—
- 8Accepted / not accepted → credit only if accepted—
Duplicate check
Three different questions, not one.
Metadata identity
ISBN, title, author, edition. Catches "same book" when the bytes differ. Cheap, first.
Exact file fingerprint
SHA-256 of the bytes, computed locally before submission. Catches the identical file — and only the identical file.
Content similarity
OCR / content fingerprint / near-duplicate detection. The only level that can tell two different scans of one book apart from two different books.
SHA-256 does not settle the question of two different scans of the same book. That is Level 3 work — future, server-side, and reviewed by people.
Optional · real, local, tiny
Try the local pre-check yourself.
The product principle: the first exact-duplicate fingerprint is calculated in your browser, before any submission. This tool does exactly that — it reads the file on your device, computes a SHA-256 fingerprint in a Web Worker so the page stays responsive, and compares it with a synthetic two-entry demo catalog. The file never leaves your device; this page has no upload at all.
Technical note: Web Crypto hashes one whole buffer at a time, so this demo caps file size (100 MB in the worker, 25 MB if the browser has to fall back to the main thread, e.g. when opened from a local file). The real product would hash in chunks — still locally. No library, no CDN.
Doctors and experts
Professional contribution is first-class.
A doctor, specialist or researcher can — in the future — contribute:
A professional credential and the correctness of a piece of material are two different things. Expert contributions go through the same review as everyone else's — credentials add a signal, not a shortcut. Credential verification is not implemented in this prototype.
XP and Trust
Trust is not a single score.
XP grows with learning and activity. Trust is a category, not a number: the future reputation model is multi-signal, and no signal is shown as one public figure like "Trust 87".
In this prototype: no real scores, no profiles, no leaderboard.
Playful, not addictive
Reward learning and verified contribution — not file volume.
| Action | What it means | Earns | When |
|---|---|---|---|
| Learn | completed a topic on the map | XP | on completion (self-check, no assessment claim) |
| Verify | found a real error in library material | Trust signal | after a reviewer confirms |
| Curate | improved metadata, mapping, classification | Trust signal | after acceptance |
| Translate | delivered a verified translation (where permitted) | Trust signal | after review |
| Contribute | provided an eligible new source or original material | Trust signal | only after pre-check → quarantine → duplicate → rights → quality → ACCEPTED |
| Review | checked another participant's material | Trust signal | after the review itself is validated |
A contribution reward exists only after a useful, confirmed result. Raw submission earns nothing.
Badges (concept)
What this game will not do
Two legally different things
Accepted ≠ trainable.
Library use — per rights profile
- search / index
- source-linked answers (RAG)
- quotation
- learning paths, quizzes, flashcards
- summaries marked as derived
MODEL_TRAINING — separate rights class
- never implied by acceptance
- its own explicit permission
- absent by default