If the Map Shows It

There is a common pattern in mesh mapping tools: draw everything you hear. A node broadcasts its neighbor list; the tool draws edges to all of them. Packets arrive with hop-path metadata; the tool connects every node in the path. The resulting map is dense and confident. It also may not be accurate.

Confidence without verification isn’t trust. It’s noise with a clean interface.

The hearing problem

When a MeshCore node transmits a packet, nearby nodes hear it and relay it forward. That process produces a lot of metadata: hop counts, path records, neighbor advertisements, signal measurements. Most of this metadata is accurate most of the time. Some of it isn’t.

Neighbor advertisement tables go stale. A node that moved six months ago still appears in another node’s neighbor list if neither has rebooted. A misconfigured node can broadcast a neighbor list that includes nodes it cannot actually reach. And in a longer-lived network, the gap between what a node claims about its neighbors and what the RF environment actually supports grows over time without anyone noticing.

A map that draws from ADVERT metadata is drawing from a node’s self-report. Self-reports are useful signals. They are not evidence.

What makes an edge verifiable

Waev draws topology edges from two sources: enrolled observers and authenticated repeaters.

An enrolled observer is a MeshCore node (or compatible device) that is connected to your MQTT broker over TCP/IP. When a packet passes through the observer’s radio range, the observer hears it, records the SNR and hop data, and publishes the observation to your broker. Waev subscribes to the broker and reads the observation. The observer is the physical evidence: it was there, it received the RF signal, it measured the SNR at that moment.

An authenticated repeater is a node that Waev can verify as a legitimate relay in your network. Its packet contributions carry provenance that can be traced. Inferred or spoofed prefixes — nodes whose identity cannot be verified against your enrolled network — are not authenticated and their edges are not drawn.

The pipeline is important to understand: Waev has no antenna. It does not receive RF directly. Everything it knows about your mesh came through an enrolled observer, over MQTT, from a real physical measurement. There is no step in that chain where Waev guesses.

What rejection means

When Waev encounters a topology source it cannot verify, it rejects it. The edge is not drawn. This happens silently — there’s no error, no placeholder, no dashed line marking the gap. The topology simply reflects only what has been verified.

This is a deliberate design choice, and it has an important implication: absence has meaning.

If a node does not appear on the Live Map, one of a small set of things is true: it is not within range of any enrolled observer; it is opted out via a privacy marker; it is offline; or its observer is not connected. Each of those is a real finding. The map is not broken; it is reporting honestly that it has no verified evidence of that node’s activity.

Compare this to a map that draws from self-reported neighbor lists. In that model, a node that went offline three weeks ago might still appear — its last-reported neighbors are still in other nodes’ tables. The map shows a ghost. The operator sees coverage that doesn’t exist, plans around it, and discovers the gap during an event rather than during a Tuesday routine check.

For a network people may lean on in an emergency, the ghost is worse than the gap.

The map’s gaps are findings

An evidence-based map’s silences are informative. A silent zone tells you: no enrolled observer covers this area. That’s actionable. Either you need to add an observer to get coverage visibility, or the zone is genuinely not reachable from your current relay infrastructure.

A zone that shows weak, intermittent edges tells you: something in that path is marginal. SNR data from the observer backs that up — a +2 dB link that was +9 dB four months ago has physically changed. The degradation is in the data, not in an assumption.

This is what we mean by evidence-based topology: the map is built from the signal record, not from a topology inference engine. For a deeper look at how the observation pipeline works — from RF packet to Live Map edge — see how Waev reads your mesh.

The position

The case for evidence-based topology is not primarily about edge cases — misconfigured nodes, stale neighbor tables, spoofed prefixes. Those exist, but they’re not the main argument.

The main argument is simpler: a map you can trust behaves differently from a map you can’t. When the map shows a link, you act on it. You plan coverage based on it. You tell a served agency it’s there. If the map is drawing from unverified self-reports, those actions rest on claims, not evidence. In a non-critical context, that’s probably fine. In an emergency, it isn’t.

Waev’s position is that the honest map — the one that says “I don’t know” rather than guessing, the one that shows gaps rather than filling them with inference — is more useful precisely because of what it refuses to draw. The gap you see is a problem you can fix. The ghost you don’t know about is a risk you’re carrying.

For more on how signal quality data informs what gets drawn and what gets flagged, see how to read SNR and hop count.

---

The map is only as good as what produced it. If you’re ready to see what your network can prove, connect at waev.app.