Skip to content

0004-content-licensing-boundary.md

Latest commit

 

History

History
68 lines (53 loc) · 3.52 KB

0004-content-licensing-boundary.md

File metadata and controls

68 lines (53 loc) · 3.52 KB
Raw

ADR-0004 — Ship an engine and an original sample; personal pools stay local

Date: 2026-07-28 · Status: Accepted · Decided by: KalevIT

Context

The previous incarnation of this project published roughly 500 questions under MIT. Two sources fed that bank, and both were checked only after the fact:

  • 299 questions tagged as derived from the LPI Learning Material. That material is licensed CC BY-NC-ND 4.0 — verified in the imprint of the official 010-160 PDF, which also repeats the notice in the footer of all 442 pages. NonCommercial and NoDerivatives are both incompatible with MIT.
  • 201 questions tagged external, traced to ThePrimoris/LPI-Essentials-Practice. That repository contains no licence file at all, which means all rights reserved. Fifteen questions were textually identical to it.

MIT does not describe what the author does with the work. It grants everyone else the right to use it commercially and to redistribute derivatives. Those are rights the author was not in a position to grant.

Options

  1. Publish anyway and hope nobody checks.
  2. Ask LPI for permission and wait before publishing anything.
  3. Publish the engine with an original sample, and have each user generate a full pool from their own copy of the freely distributed material.

Advice received

Option 3, with option 2 available later if the pool ever needs to be distributed. The reasoning rests on three distinctions the previous version had collapsed into one:

  • Facts are not copyrightable. That UID 0 is root, that $? holds an exit status — these belong to nobody. Copyright protects expression.
  • The exam objectives are not the learning material. LPI publishes the objectives as a specification so that candidates and educators can prepare. Third-party certification books have been written against published objectives for decades.
  • NoDerivatives restricts distribution, not private use. Generating a personal question pool from your own copy of a freely given PDF, to study with, is not distribution.

Decision

Option 3.

  • The repository ships an engine plus 40 original questions written from the published objectives and general technical knowledge — one per official weight point, exactly one full weighted exam, enough to demonstrate the engine and honest about not being a study bank.
  • Personal pools live in data/questions.personal.js, which is git-ignored. core.js uses it automatically when present.
  • Questions carry a provenance tag: obj from the objectives, own from personal experience, mat derived from the learning material. The CI validator fails the build if a mat question appears in a committed bank.
  • Code stays MIT. Content terms are stated separately in LICENSE-CONTENT.md.
  • The 15 questions traced to the unlicensed repository are not carried over in any form.

Consequences

  • A fresh clone has 40 questions, not 500. Anyone wanting a real study bank follows docs/QUESTION-SOURCING.md and builds their own.
  • The project's centre of gravity moves from a simulator for one exam to an engine that turns official material into a simulator for any LPI exam. The licensing constraint pushed the architecture where it was already heading.
  • Nothing in the repository depends on anyone else's copyright.
  • This is not legal advice and its author is not a lawyer. The material's own imprint invites questions at learning@lpi.org; that route stays open if the position ever needs confirming, particularly for commercial use.