Three altitudes
Reliability analysts already know machine reliability. Developer tooling taught a second altitude: ecosystem / DevX reliability — can dependents keep working when symbols and packages move? The public web forces a third: internet reliability — can shared references stay usable when labels, hosts, and paths churn?
flowchart TB
subgraph internet [Internet reliability]
direction TB
subgraph eco [Developer-ecosystem / DevX reliability]
direction TB
subgraph machine [Machine reliability]
M[Process · disk · host uptime]
end
E[Packages · docs xrefs · agent trees]
end
I[Cool URIs · citations · shared memory]
end
machine --> eco --> internet
Nested, not rivals
| Altitude | Failure people feel | Typical fix plane |
|---|---|---|
| Machine | Crash, disk full, node down | Ops, HA, backups |
| Developer ecosystem | Broken import, wrong docs tree, agent archaeology | Graphs, CONSUMERS.md, package reverse-deps |
| Internet | Link rot, rename cascades, “half the org’s bookmarks died” | Identity stack + absorb layer + durable citation norms |
A site can have five-nines uptime and still be unreliable as a reference if every rename orphans consumers.
Branding note
DevCentr names the practitioner surface inside broader system / digital architecture. UI/UX (HCI Nerdz) is architecture instantiated in spatial and sensory modalities — essential, not the sole umbrella for internet-scale reference identity. Keep product SPE Software Architecture for ship checklists; keep this altitude for the public reference plane.
Next
-
Reference reliability — the URL-rot slice
-
Identity stack — how the altitudes share one treatment stack