research: correct verified license-text errors in comparison corpus #20

Activity

1. zackees

   added a parent issue on Aug 25, 2026

   • audit: adversarially review the paralegal research corpus #19
2. zackees commented on Aug 25, 2026

   

   zackees

   on Aug 25, 2026

   MemberAuthor

   More actions

   Sol adversarial review - 2026-08-25

   Verdict: VALID. Leave open. I independently tried to falsify each operative-text claim against the current steward/official-host texts. All four core mismatch clusters are confirmed on default-branch commit 40ee710; the relevant files were introduced by PR #18, no later commit changes them, no open PR addresses them, and the issue/PR search found no true duplicate or completed resolution. This is preliminary research for attorney review, not legal advice.

   Confirmed claims and narrowed corrections

   • RPL-1.5: OSI's approved RPL-1.5 text confirms §1.2 defines “Deploy”; §6.1 requires Source Code for deployed Extensions to be available within one month and retained for at least 12 months after deployment ceases; §6.4(c) separately requires community notification within one month. §12.1 is a 30-day cure measured from awareness of the breach, not a first-notice rule; §12.2 is patent-action termination; §13.8 selects Colorado law, AAA arbitration, and Adams County venue. AUTH-0018 currently miscites §§1.12/3.2, attributes cure to §8, and says California/San Francisco; PR-0002 incorrectly says source is public “at deployment,” uses a first-breach/notice formulation, and repeats California/San Francisco. Narrowing: the corrected disclosure statement should track §6.1's “Extensions that You Deploy,” rather than broadly saying all Licensed Software and Extensions.
   • OSL-3.0: OSI's approved OSL-3.0 text §11 says an action or suit may be brought only in courts where the Licensor resides or conducts its primary business, under that jurisdiction's law (excluding conflicts rules). AUTH-0017 and PR-0002 therefore cannot say forum/law is unstated. Narrowing: this is a location-dependent rule, not a fixed named jurisdiction.
   • CERN-OHL-S-2.0: CERN says the texts linked from its licence page are definitive; the definitive CERN-OHL-S v2 text §§8.4-8.5 provides immediate termination, reinstatement from each Licensor after all breaches cease unless that Licensor gave notice while the licensee remained in breach and material compliance was not achieved within 30 days, and permanent termination for a repeated breach after notice and reinstatement. AUTH-0139 and PR-0002 incorrectly say no general cure/reinstatement is stated. Preserve the existing open-hardware/design-source scope warning.
   • MPL-2.0: Mozilla's official MPL-2.0 text §5.1 provides automatic termination and per-Contributor provisional/ongoing reinstatement: ongoing reinstatement follows if no notice arrives before 60 days after restored compliance, or after first notice from that Contributor if compliance occurs within 30 days. §8 instead governs litigation/forum/law. PR-0011 repeatedly and incorrectly points its cure/reinstatement discussion and scenario to §8. AUTH-0013 and PR-0002 already point generally to §5, but should identify §5.1 and state the mechanism precisely.

   Missing acceptance detail

   Add explicit checks for RPL §§1.2 and 6.4(c), its 12-month post-deployment availability duration, and the §6.1 Extensions-only wording. For CERN, preserve “from any Licensor,” notice while still in breach, and “in all material respects.” For MPL, preserve the per-Contributor distinction between cessation/provisional reinstatement, the 60-day no-notice path, and the 30-day first-notice path. A corpus search should verify that LEGAL-REVIEW.md and README.md retain their already-correct MPL §5.1 references; PROPOSITION-MAP.md presently has no contradictory license-specific section claim requiring a substantive rewrite.

   Remaining attorney-review questions

   1. Should the normalized matrix split pre-termination cure, provisional reinstatement after cessation, and permanent/ongoing reinstatement into separate fields?
   2. Should dynamic forum/law rules and arbitration clauses be normalized separately from a simple named-jurisdiction field?
   3. For RPL, should the comparison rely strictly on operative §§6.1/6.4(c) rather than the preamble's broader “when you Deploy” shorthand?

   The canonical research records and all legal-review statuses should remain legal_review: pending; this issue does not authorize a LICENSE change.