33 lines
1.3 KiB
Markdown
33 lines
1.3 KiB
Markdown
# ADR 0001: Treat LixelGO as unavailable
|
|
|
|
Status: accepted.
|
|
|
|
## Context
|
|
|
|
The actual stand consists of one K1, one MacBook and one ordinary TP-Link
|
|
Deco/mesh router. There is no phone, LixelGO reference session, Linux host,
|
|
dedicated capture router or public SDK.
|
|
|
|
The supplied design documents include several app-dependent experiments and
|
|
assume that provisioning payloads can be recovered by comparing LixelGO BLE
|
|
writes. That evidence source does not exist.
|
|
|
|
## Decision
|
|
|
|
- LixelGO is not a temporary dependency, broker or test oracle.
|
|
- App-dependent experiments remain only as historical reference.
|
|
- The active critical path is physical baseline, passive BLE, remembered Wi-Fi
|
|
or device AP, evidence-led provisioning, application-session discovery, then
|
|
stream capture and decoding.
|
|
- A fully opaque provisioning protocol with no evidence source is reported as a
|
|
hard black-box blocker rather than approached with random writes.
|
|
- The existing router is used without changes for the first observation gate.
|
|
|
|
## Consequences
|
|
|
|
The project may reach a legitimate BLOCKED result before receiving live data.
|
|
This is still a successful pre-production conclusion because it identifies the
|
|
exact missing interface and prevents premature decoder work or risky device
|
|
mutation.
|
|
|