← Garden Patch Home · Inquiries
IFP-5 defines four authentication levels as discrete stages in a progressive trust model. Each level represents a distinct verification mechanism:
| Level | Mechanism | Jump from Previous |
|---|---|---|
| 0 | Shared secret (introduction token) | — |
| 1 | Public-key signature | Symmetric → asymmetric crypto |
| 2 | Key verified via identity document | Self-asserted → externally resolvable |
| 3 | Key bound to DID | Domain-specific → decentralized identity |
The jumps between levels are large. Are there trust-relevant distinctions being collapsed? Are four stages the right granularity?
The Level 0 → 1 jump is the largest. Moving from a shared secret to public-key signatures changes the entire trust model — from “someone we both know introduced us” to “I can verify your signature independently.” This single jump may be doing too much work.
Level 3 depends on DID infrastructure that does not yet exist at scale. If Level 3 is aspirational rather than practical, IFP effectively has three usable levels. Is that enough?
Christopher Allen’s progressive trust framework describes a richer spectrum than four discrete stages. The question is whether IFP’s four-level discretization loses important trust distinctions that the continuous model preserves.
The minimum viable architecture question applies. Are four levels a load-bearing architectural decision (the right boundaries, worth committing to early) or a tactical choice (the specific boundaries could be refined without reshaping the protocol)?