Changelog: Logos Storage in Testnet v0.2
Record of the adversarial review applied to the storage draft. Process: three independent critics (research advocate checking claims against the roadmap repo and docs.logos.co, narrative advocate, adversary) attack the draft in parallel; a synthesizer revises under the author’s voice profile; repeat. Two rounds run; effectively converged in round 2.
Issues found and fixed
| # | Found by | Issue | Fix |
|---|---|---|---|
| 1 | Research advocate, R1 | Draft said NAT traversal “is landing in v0.2.1” — no roadmap file mentions a v0.2.1 release; the label existed only in the comms brief | Reframed to what sources support: package complete (TCP hole-punching transfers landed late June per storage/updates/2026-06-22.md), absent from v0.2 release notes, expected in a follow-up release |
| 2 | Research advocate + adversary, R1 (independently) | “A network observer no longer gets a map of who wants what” overclaimed the DHT-over-mix PoC and contradicted the draft’s own later admission that retrieval is watchable | Rescoped: PoC, lookup path only, passive observer; retrieval explicitly still visible |
| 3 | Adversary, R1 | The draft conceded the shipped product looks like IPFS but never answered the “why not IPFS?” comparison its audience would make | New paragraph: integration (Logos Core module, Status archival-storage integration) is the shipped answer, the privacy program the promised one, with the explicit concession that “IPFS with extra steps” becomes a fair review if the program slips |
| 4 | Adversary, R1 | Content addressing oversold as censorship resistance; availability ignored | Added the caveat: takedown-resistance is content addressing plus replication, which also gives the NAT traversal section its stakes |
| 5 | Research advocate, R1 | ”Late 2026” imprecise; “same mix network” wrong (blockchain uses Blend, not the mix protocol); “most storage networks simply ignore” unsupported sweep; “nothing is a special child anymore” too broad | October 2026 per privacy-preserving-filesharing.md; “same mix protocol”; property claim about vanilla Kademlia DHTs; narrowed to the two components the release notes support |
| 6 | Research advocate, R1 | ”Architectural improvements” to the block protocol undersold it: the old protocol was known-broken; v0.2 ships a new one | Rewritten as “a new block protocol” with the roadmap’s own “broken in many ways” framing |
| 7 | Narrative advocate, R1 | ”The honest part” opened by announcing honesty (voice-profile violation); dense privacy-program paragraph; vague operator ask | Section opens with its strongest sentence; paragraph split; ask made concrete (scrape OpenMetrics, hammer large transfers) |
Critiques rejected (with reasons)
- Cutting the content-addressing section to reach the release notes faster (narrative advocate, R1): rejected; the section now carries the availability caveat the adversary required.
- Softening the lede ultimatum (“serve the frontend from decentralised storage or stop claiming the adjective”) (adversary, R2): retained as deliberate stance; the voice profile mandates taking positions, and the adversary itself judged it rhetorical exposure rather than a logical flaw.
Known weaknesses (unresolved by design)
- The lede ultimatum is attackable by readers who count ENS+IPFS gateways as adequate; it is an opinion and framed as one.
- Cross-article links point at content-tracker draft filenames until published URLs exist (publish-time TODO in the draft).
- “Expect it in a follow-up release” (NAT traversal) is inference from package completion, not a sourced commitment.
Ground truth consulted: roadmap/content/testnets/v02.md, v02-release.md, storage/roadmap/filesharing.md, privacy-preserving-filesharing.md, storage/roadmap/tracks/v0.1-filesharing.md, storage/updates/2026-06-22.md, 2026-06-29.md; docs.logos.co URL index (llms.txt).