Rx
rx-controls-suite · Reactive Programming for Scientific Control Systems
Conference Talks
Same ReactiveX idioms.
Every control system.
poll  ·  zip  ·  sliding average  ·  backpressure  ·  fluent pipelines
choose a talk to begin
Start Here
Reactive Programming,
Explained With Marbles
Primer Not a talk
Old style vs. reactive — same logic, side by side
Animated marble diagram — watch the pipeline run
Live dashboard — the real thing, not just a diagram
Open primer →
Tango Users Meeting
Reactive Programming
for Tango Controls
Java jbang RxJava3 C++ simulator
zip — correlated attribute reads
buffer + distinctUntilChanged — sliding avg
merge — alarm fan-in
Open talk →
EPICS Collaboration Meeting
Reactive Programming
for EPICS Controls
Python uv RxPY v4 caproto
rx.zip — correlated PV reads
buffer_with_count + distinct_until_changed
merge — alarm fan-in
Open talk →
Tango Users Meeting — Python Edition
Reactive Programming
for Tango Controls (Python)
Python uv RxPY v4 PyTango
rx.zip — correlated attribute reads
buffer_with_count + distinct_until_changed
merge — alarm fan-in · execute_command
Open talk →
DESY · TINE Middleware
Reactive Programming
for TINE Controls
Java jbang RxJava3 TINE Java API
zip — correlated property reads
buffer + distinctUntilChanged — monitor stream
merge — alarm fan-in · no commands
Open talk →
Tango Users Meeting — C++ Edition
Reactive Programming
for Tango Controls (C++)
C++17 CMake RxCpp cppTango
zip — correlated attribute reads
buffer + distinct_until_changed — sliding avg
merge — alarm fan-in · execute_command
Open talk →
EPICS Collaboration Meeting — C++ Edition
Reactive Programming
for EPICS Controls (C++)
C++17 CMake RxCpp PVXS
zip — correlated PV reads
buffer + distinct_until_changed
merge — alarm fan-in · no commands
Open talk →
Combined Demo — Capstone
Storage Ring × Beamline
One Reactive Pipeline
Python rxtango rxepics Docker
rx.zip (cross-system) — EPICS + Tango, atomic frame
share + filter + take_until — beam-loss · abort
distinct_until_changed — shutter supervisor
Open talk →
Workflow Orchestration
Same Scan, New Conductor
Orchestrating With Prefect
Prefect Reused core Auto pause/resume
Same guarded scan — reused, not rewritten
Two scenarios — happy path vs. beam-loss pause/resume
For managers & scientists — no code required
Open talk →
Workflow Orchestration — No-Python Edition
Same Scan, Visual Graph
Orchestrating With n8n
n8n HTTP + SSE Refinement loop
A cycle, not a DAG — re-acquire what came out LOW
The scan as a REST resource — what that costs and buys
Wait node + auto-resume — beam loss, no operator click
Open talk →