Add static network

SPEC.md

@@ -344,6 +344,70 @@ static_assert(
 Multi-output functions must return `std::tuple<...>`. Single return accepted as-is.
 `void` return = sink node (no output ports).
 
+### Node identity: `Label` and `UniqueTag` NTTPs
+
+Two problems arise when using `Node` as a compile-time graph vertex:
+
+1. **Debug names** — without a label the web UI and `print_diagnostics` fall back to
+   `"node[0]"` placeholders. `set_name()` provides a runtime name for the runtime
+   `Network`, but `StaticNetwork` has no `add("name", node)` call to attach one.
+
+2. **Same-function collision** — two nodes wrapping the same function (e.g. two
+   `make_node<blur_func>`) have the same type. The `StaticNetwork` topo sort and fanout
+   detection use node types as graph vertices, so it cannot distinguish them, causing
+   infinite recursion in the DFS.
+
+Both are solved by two additional NTTPs on `Node`:
+
+```cpp
+template<
+    auto Func,
+    typename InputTag  = in<>,
+    typename LatchTag  = latch<>,
+    typename OutputTag = out<>,
+    fixed_string Label = "",       // human-readable name; shown in diagnostics/web UI
+    std::size_t UniqueTag = 0      // collision-breaker; must differ between any two
+>                                  // same-Func nodes in one make_network() call
+class Node : public INode { ... };
+```
+
+Both have defaults so all existing code (`make_node<blur_func>(5)`) compiles unchanged.
+
+**Factory syntax extension:**
+
+```cpp
+// Label only (UniqueTag defaults to 0)
+auto blur = make_node<blur_func, "blur">(8);
+
+// Both label and unique tag — required when the same function is used twice
+auto preblur  = make_node<blur_func, "preblur",  0>(8);
+auto postblur = make_node<blur_func, "postblur", 1>(8);
+```
+
+**Duplicate-tag detection** — `make_network()` checks at compile time that no two nodes
+in the edge pack share the same `(Func, UniqueTag)` pair. This fires a readable
+`static_assert` at the call site before any thread is started:
+
+```
+static_assert(!has_duplicate_tags_v<AllUserNodes>,
+    "make_network: two nodes have identical (Func, UniqueTag) — increment UniqueTag "
+    "on one of them to make them distinct");
+```
+
+**Label availability** — `Node` exposes the label as a `static constexpr` so
+`StaticNetwork` can read it at compile time for diagnostics:
+
+```cpp
+static constexpr std::string_view label() { return Label.view(); }
+```
+
+The web debug UI and `print_diagnostics` use `label()` when non-empty, falling back to
+`"node[<UniqueTag>]"` for unlabelled nodes. The runtime `Network` continues to use the
+string passed to `add("name", node)` — `Label` is orthogonal to that mechanism.
+
+**`ObjectNode`** gains the same two NTTPs with the same defaults. The `make_node(obj, ...)`
+overloads are extended identically.
+
 ---
 
 ## Component 4a — Latched Input Ports
@@ -1065,6 +1129,231 @@ net.start();
 
 ---
 
+## Component 9 — `static_network.hpp`: Compile-time Graph Builder
+
+### Motivation
+
+The runtime `Network` builder has two limitations that only a compile-time graph can fix:
+
+1. **Fan-out `N` is unknowable at the first `connect()` call.** A `FanoutNode<T, N>` requires
+   `N` as a template parameter. With runtime `connect()`, the network has seen only one edge
+   when the first call arrives; it cannot know how many more will follow for that port. Auto-
+   inserting the right `FanoutNode<T, N>` requires seeing the complete edge list at once —
+   which is only possible if the edge list is a type.
+
+2. **Start/stop goes through virtual dispatch.** `Network` stores `INode*` and calls virtual
+   `start()`/`stop()`. With a typed node tuple the compiler sees the concrete types and can
+   inline or devirtualise. This matters at startup/shutdown, not in the hot path — but it is
+   avoidable overhead.
+
+The runtime `Network` is **not removed**. It remains the right choice for Python graphs,
+sub-networks embedded in dynamic topologies, and any case where the graph shape is not known
+until runtime. `StaticNetwork` is an additional builder for the common case where the full
+C++ topology is known at compile time.
+
+### API
+
+```cpp
+// edge() constructs a typed edge descriptor from two port handles.
+// All type information (SrcNode, SrcIdx, DstNode, DstIdx) is in the return type.
+template<typename SrcNode, std::size_t SrcIdx,
+         typename DstNode, std::size_t DstIdx>
+auto edge(OutputPort<SrcNode, SrcIdx>, InputPort<DstNode, DstIdx>)
+    -> Edge<SrcNode, SrcIdx, DstNode, DstIdx>;
+
+// make_network() accepts all edges as a variadic pack.
+// It deduces the full topology, auto-inserts FanoutNodes where needed,
+// wires all channels, and returns a StaticNetwork owning the fanout nodes.
+// User nodes are held by reference (non-owning), same lifetime contract as Network.
+template<typename... Edges>
+auto make_network(Edges&&... edges) -> StaticNetwork<...>;
+```
+
+Usage:
+
+```cpp
+auto src    = make_node<produce>(8);
+auto blur   = make_node<blur_func,   in<"img">>(8);
+auto detect = make_node<detect_func, in<"img">>(8);
+auto sink   = make_node<display>(8);
+
+// src:output<0> feeds both blur and detect — fan-out is auto-inserted
+auto net = make_network(
+    edge(src.output<0>(),    blur.input<0>()),
+    edge(src.output<0>(),    detect.input<0>()),   // same source port
+    edge(blur.output<0>(),   sink.input<0>()),
+    edge(detect.output<0>(), sink.input<1>())
+);
+net.start();
+// ...
+net.stop();
+```
+
+No `add()`, no `build()`, no string names. The graph is fully wired in the `make_network`
+call. `start()` and `stop()` are non-virtual tuple traversals.
+
+### Edge type
+
+```cpp
+// Carries references to the two endpoint nodes. Stores no data beyond that.
+template<typename SrcNode, std::size_t SrcIdx,
+         typename DstNode, std::size_t DstIdx>
+struct Edge {
+    SrcNode& src;
+    DstNode& dst;
+};
+```
+
+### Fan-out detection metafunction
+
+`make_network` receives `Edge<...>` types as a pack. Before wiring, a metafunction scans
+the pack for output ports with more than one downstream edge:
+
+```
+fanout_groups<Edges...>
+```
+
+This is a compile-time multimap: keys are `(SrcNode type, SrcIdx)`, values are the list of
+destination `(DstNode type, DstIdx)` pairs sharing that key.
+
+For each key with N > 1 destinations:
+- Compute `T = std::tuple_element_t<SrcIdx, SrcNode::return_tuple>`
+- Synthesise a `FanoutNode<T, N>` — call it `F`
+- Replace the N original edges with:
+  - one edge: `src:SrcIdx → F:input<0>`
+  - N edges:  `F:output<0..N-1> → original dst:DstIdx`
+
+For keys with N == 1 the edge is kept as-is.
+
+The result is an expanded edge list with all fan-outs made explicit, and a list of
+`FanoutNode<T, N>` types that need to be instantiated.
+
+### `StaticNetwork` structure
+
+```cpp
+template<typename UserNodeRefs,   // tuple of T& for each user node
+         typename FanoutStorage,  // tuple of FanoutNode<T,N> owned by the network
+         typename TopoOrder>      // index_sequence encoding start/stop order
+class StaticNetwork : public INode {
+public:
+    void start();   // std::apply over TopoOrder — no virtual dispatch, no map lookup
+    void stop();    // reverse of TopoOrder
+
+    bool running() const;
+
+    // Diagnostics — iterates typed tuples; same NodeSnapshot / ChannelSnapshot output
+    // as Network, compatible with print_diagnostics and the web debug UI.
+    void print_diagnostics(std::ostream& = std::cerr) const;
+
+    // StaticNetwork is itself an INode, so it can be embedded in a runtime Network
+    // via net.add("stage", static_net) exactly like any other node.
+    void set_name(std::string) override;
+    const NodeStats& stats() const override;
+    NodeSnapshot node_snapshot(const std::string&, double) const override;
+
+private:
+    FanoutStorage fanouts_;   // owns the auto-generated FanoutNode instances
+    // User nodes held by reference — same non-owning contract as Network
+};
+```
+
+`FanoutStorage` is a `std::tuple<FanoutNode<T0,N0>, FanoutNode<T1,N1>, ...>` with one
+element per auto-inserted fanout. It is owned by the `StaticNetwork` and lives as long as
+the network does — which satisfies the channel lifetime contract (channels are owned by their
+consumer, and the fanout node is the consumer of the upstream output).
+
+### Cycle detection
+
+With the full edge list as a type pack, cycle detection is a `static_assert` rather than a
+runtime exception. A compile-time DFS over the expanded edge list fires a readable assertion
+at the `make_network` call site:
+
+```
+static_assert(!has_cycle_v<ExpandedEdges...>,
+    "make_network: graph contains a directed cycle");
+```
+
+`NetworkCycleError` is no longer needed for `StaticNetwork` — the cycle is caught before any
+object is constructed.
+
+### Topological order
+
+The same compile-time DFS produces a topological ordering as an `std::index_sequence` over
+the node tuple. `start()` iterates it forward, `stop()` iterates it in reverse. No runtime
+sort, no `std::vector<std::string>`.
+
+### Node labels for diagnostics / web debug
+
+Labels come directly from the `Label` NTTP on each `Node` type — no separate annotation
+on `edge()` is needed. `StaticNetwork` reads `NodeType::label()` at compile time for each
+vertex in the topo order and stores the result as a `std::string_view` array at
+construction time. Zero runtime overhead: the label is a compile-time string literal.
+
+```cpp
+auto src    = make_node<produce,    "src"   >(8);
+auto blur   = make_node<blur_func,  "blur"  >(8);
+auto detect = make_node<detect_func,"detect">(8);
+
+auto net = make_network(
+    edge(src.output<0>(),    blur.input<0>()),
+    edge(src.output<0>(),    detect.input<0>())
+);
+// web UI shows nodes named "src", "blur", "detect"
+// auto-inserted FanoutNode is labelled "fanout[src:0]"
+```
+
+Unlabelled nodes (`Label == ""`) fall back to `"node[<UniqueTag>]"` in diagnostics.
+Auto-inserted fanout nodes are labelled `"fanout[<src_label>:<SrcIdx>]"` automatically.
+
+### Wiring sequence in `make_network`
+
+All wiring happens in the `make_network` constructor body — no `build()` call needed:
+
+1. Instantiate `FanoutStorage` (default-construct each `FanoutNode<T,N>`).
+2. For each expanded edge (in topological order):
+   - Call `src.set_output_channel<SrcIdx>(&dst.input_channel<DstIdx>())`.
+3. Return the `StaticNetwork`.
+
+Channel pointers are set once and never changed. No dynamic allocation after construction.
+
+### What is eliminated vs `Network`
+
+| `Network` (runtime) | `StaticNetwork` (compile-time) |
+|---|---|
+| `std::map<string, INode*>` | typed `std::tuple` of references |
+| Runtime DFS + `NetworkCycleError` | `static_assert` at `make_network` call site |
+| Virtual `start()`/`stop()` per node | `std::apply` over typed tuple |
+| Explicit `make_fanout<T,N>` | auto-inserted from edge pack |
+| `connected_outputs_` duplicate check | structural impossibility — no duplicate edge can produce two `set_output_channel` calls |
+| `build()` step | no build step — wired in constructor |
+
+The hot path (per-item `pop` → `push` in each node thread) is identical in both cases.
+
+### Compatibility
+
+- `StaticNetwork` implements `INode`, so it can be registered inside a runtime `Network`
+  via `net.add("name", static_net)` — enabling mixed static/dynamic graphs.
+- All existing node types (`Node`, `ObjectNode`, `FanoutNode`, `MainThreadNode`) work
+  unchanged as vertices in a `StaticNetwork`.
+- The Python `PyNetwork` is unaffected — it remains runtime-only.
+
+### File layout addition
+
+```
+include/kpn/
+    static_network.hpp    # Edge<>, make_network(), StaticNetwork<>
+    tmp/
+        fanout_groups.hpp  # fanout_groups<Edges...> metafunction
+        topo_sort.hpp      # compile-time DFS + cycle check
+        repeat_tuple.hpp   # repeat_tuple_t<T,N> (moved from fanout.hpp)
+```
+
+`fanout.hpp` keeps `FanoutNode<T,N>` and `make_fanout<T,N>` for users who want to wire
+fanouts explicitly in a runtime `Network`. `static_network.hpp` uses `FanoutNode` internally
+but the user never calls `make_fanout` when using `make_network`.
+
+---
+
 ## Resolved Design Decisions
 
 All major design questions are now closed:
@@ -1079,3 +1368,6 @@ All major design questions are now closed:
 | `make_py_network` | Pure C++ template; nanobind module recompilation is the registration step |
 | GIL strategy | Acquire per Python callback; release while blocking on channel ops |
 | Mixed-rate inputs | `latch<>` tag for ports that reuse last-seen value; blocks only on first fire; node fires at rate of `in<>` ports |
+| Fan-out | Explicit `FanoutNode<T,N>` for runtime `Network`; auto-inserted by `make_network()` for `StaticNetwork` |
+| Static vs runtime graph | Both coexist; `StaticNetwork` for C++ graphs known at compile time, `Network` for Python/dynamic graphs; `StaticNetwork` implements `INode` so it embeds in `Network` |
+| Node identity in static graphs | `Label` NTTP (human name for diagnostics) + `UniqueTag` NTTP (collision-breaker for same-Func nodes); both default to `""` / `0` so existing code is unaffected |

examples/10_static_hello_pipeline/main.cpp

@@ -0,0 +1,34 @@
+// Example 10 — Static Hello Pipeline
+//
+// The same linear pipeline as example 01, built with make_network() instead
+// of the runtime Network builder. The topology is fully known at compile time:
+// cycle detection is a static_assert, no build() step is needed, and start/stop
+// require no virtual dispatch through a string-keyed node map.
+//
+//  [produce] --int--> [double_it] --int--> [print_it]
+
+#include <kpn/kpn.hpp>
+#include <chrono>
+#include <iostream>
+#include <thread>
+
+static int  produce()        { return 42; }
+static int  double_it(int x) { return x * 2; }
+static void print_it(int x)  { std::cout << "result: " << x << '\n'; }
+
+int main() {
+    using namespace kpn;
+
+    auto src  = make_node<produce,   "src" >(5);
+    auto dbl  = make_node<double_it, "dbl" >(5);
+    auto sink = make_node<print_it,  "sink">(5);
+
+    auto net = make_network(
+        edge(src.output<0>(),  dbl.input<0>()),
+        edge(dbl.output<0>(),  sink.input<0>())
+    );
+
+    net.start();
+    std::this_thread::sleep_for(std::chrono::milliseconds(100));
+    net.stop();
+}

examples/11_static_fanout/main.cpp

@@ -0,0 +1,55 @@
+// Example 11 — Static Fan-Out (automatic FanoutNode insertion)
+//
+// Two consumers read from the same output port of [generate]. With the runtime
+// Network builder this would require an explicit make_fanout<>. With make_network()
+// the duplicate source port is detected at compile time and a FanoutNode<string,2>
+// is inserted automatically — the user just writes two edges from the same port.
+//
+//                             +--> [print_key]
+//  [generate] --string--> [fan]
+//                             +--> [print_upper]
+//
+// The FanoutNode<string,2> is owned by the StaticNetwork and invisible to the user.
+
+#include <kpn/kpn.hpp>
+#include <algorithm>
+#include <chrono>
+#include <iostream>
+#include <string>
+#include <thread>
+
+static int gen_index = 0;
+
+static std::string generate() {
+    static const char* words[] = {"hello", "kpn", "fanout", "static", "network"};
+    std::this_thread::sleep_for(std::chrono::milliseconds(80));
+    return words[gen_index++ % 5];
+}
+
+static void print_lower(std::string s) {
+    std::cout << "lower:  " << s << '\n';
+}
+
+static void print_upper(std::string s) {
+    std::transform(s.begin(), s.end(), s.begin(), ::toupper);
+    std::cout << "upper:  " << s << '\n';
+}
+
+int main() {
+    using namespace kpn;
+
+    auto gen   = make_node<generate,    "gen"  >(5);
+    auto lower = make_node<print_lower, "lower">(5);
+    auto upper = make_node<print_upper, "upper">(5);
+
+    // Two edges from gen.output<0>() — make_network() detects the fan-out
+    // and inserts FanoutNode<std::string, 2> automatically.
+    auto net = make_network(
+        edge(gen.output<0>(),  lower.input<0>()),
+        edge(gen.output<0>(),  upper.input<0>())
+    );
+
+    net.start();
+    std::this_thread::sleep_for(std::chrono::milliseconds(500));
+    net.stop();
+}

examples/12_static_cellshade/main.cpp

@@ -0,0 +1,217 @@
+// Example 12 — Static Cell-Shading Pipeline with Auto Fan-Out
+//
+// The same cell-shading effect as example 09, rebuilt with make_network().
+//
+// Key differences from example 09:
+//
+//  1. Fan-out is automatic. The edge detector output feeds both the compositing
+//     node and the debug display window. In example 09 this required composite()
+//     to re-output the edge mask as a second return value (a workaround). Here
+//     make_network() detects the duplicate source port and inserts
+//     FanoutNode<cv::Mat, 2> automatically — composite() is a clean single-output
+//     node.
+//
+//  2. No add()/connect()/build() ceremony. The full topology is expressed once
+//     in the make_network() call. Cycle detection and duplicate-tag checking are
+//     compile-time static_asserts.
+//
+//  3. Every node has a Label NTTP so the web debug UI shows real names.
+//
+// Topology:
+//
+//  [capture] --colour--> [quant] ──────────────────────────────> [comp] --> [display_composite]
+//  [capture] --grey----> [to_gray] --> [edges] --edges--(fan)--> [comp]
+//                                               --edges----------> [display_edges]   ← auto-fanout
+//
+// Note: capture returns std::tuple<cv::Mat, cv::Mat> (colour, grey).
+//       The two outputs are separate ports routed independently.
+
+#include <kpn/kpn.hpp>
+#include <opencv2/core.hpp>
+#include <opencv2/imgproc.hpp>
+#include <opencv2/highgui.hpp>
+#include <opencv2/videoio.hpp>
+#include <iostream>
+#include <tuple>
+#include <thread>
+#include <chrono>
+
+// ── Gradient base for synthetic pattern ───────────────────────────────────────
+
+static cv::Mat make_gradient(int W, int H) {
+    cv::Mat xr(H, W, CV_8UC1), yg(H, W, CV_8UC1), b(H, W, CV_8UC1, cv::Scalar(128));
+    for (int x = 0; x < W; ++x) xr.col(x).setTo(x * 255 / W);
+    for (int y = 0; y < H; ++y) yg.row(y).setTo(y * 255 / H);
+    cv::Mat channels[3] = {b, yg, xr};
+    cv::Mat grad;
+    cv::merge(channels, 3, grad);
+    return grad;
+}
+
+// ── Pipeline functions ────────────────────────────────────────────────────────
+
+static std::tuple<cv::Mat, cv::Mat> capture() {
+    constexpr int W = 640, H = 480;
+    static cv::VideoCapture cap;
+    static bool opened = false;
+    if (!opened) {
+        opened = true;
+        cap.open(0, cv::CAP_V4L2);
+        if (cap.isOpened()) {
+            cap.set(cv::CAP_PROP_FRAME_WIDTH,  W);
+            cap.set(cv::CAP_PROP_FRAME_HEIGHT, H);
+        } else {
+            std::cerr << "[capture] no webcam — using synthetic animated pattern\n";
+        }
+    }
+
+    cv::Mat frame;
+    if (cap.isOpened()) {
+        auto t0 = std::chrono::steady_clock::now();
+        cap >> frame;
+        auto elapsed = std::chrono::steady_clock::now() - t0;
+        if (elapsed < std::chrono::milliseconds(20))
+            std::this_thread::sleep_for(std::chrono::milliseconds(33) - elapsed);
+        if (frame.empty()) frame = cv::Mat::zeros(H, W, CV_8UC3);
+    } else {
+        static int tick = 0;
+        static cv::Mat grad = make_gradient(W, H);
+        ++tick;
+        frame = grad.clone();
+        int r = 150 + (tick % 80) * 4;
+        cv::circle(frame, {W/2,   H/2},   r,     {255, 200,   0}, -1);
+        cv::circle(frame, {W/2,   H/2},   r / 2, {  0, 128, 255}, -1);
+        cv::circle(frame, {W*2/5, H*2/5}, r / 3, {200,   0, 200}, -1);
+        std::this_thread::sleep_for(std::chrono::milliseconds(33));
+    }
+    return {frame.clone(), frame.clone()};
+}
+
+static cv::Mat to_gray(cv::Mat bgr) {
+    cv::Mat gray;
+    cv::cvtColor(bgr, gray, cv::COLOR_BGR2GRAY);
+    return gray;
+}
+
+static cv::Mat edges_fn(cv::Mat gray) {
+    cv::Mat blurred, mask;
+    cv::GaussianBlur(gray, blurred, {5, 5}, 0);
+    cv::Canny(blurred, mask, 50, 150);
+    return mask;
+}
+
+static cv::Mat quantise(cv::Mat bgr) {
+    constexpr int levels = 4;
+    constexpr double step = 256.0 / levels;
+    static const cv::Mat lut = []() {
+        cv::Mat l(1, 256, CV_8UC1);
+        for (int i = 0; i < 256; ++i)
+            l.at<uchar>(i) = cv::saturate_cast<uchar>(
+                std::floor(i / step) * step + step / 2.0);
+        return l;
+    }();
+    cv::Mat out;
+    cv::LUT(bgr, lut, out);
+    return out;
+}
+
+// Clean single-output composite — no longer needs to pass edges through.
+static cv::Mat composite(cv::Mat edge_mask, cv::Mat colour) {
+    cv::Mat result = colour.clone();
+    result.setTo(cv::Scalar(0, 0, 0), edge_mask);
+    return result;
+}
+
+// ── Display nodes ─────────────────────────────────────────────────────────────
+//
+// Two separate MainThreadNode subclasses — one for the composited result,
+// one for the raw edge mask. Each runs on the main thread via step().
+// The fan-out from [edges] to both consumers is inserted automatically by
+// make_network().
+
+class DisplayComposite : public kpn::MainThreadNode<DisplayComposite,
+                                                    kpn::in<"composite">,
+                                                    cv::Mat> {
+public:
+    // Label and unique_tag for StaticNetwork identity
+    static constexpr std::string_view label()    { return "display_composite"; }
+    static constexpr std::size_t      unique_tag = 0;
+
+    DisplayComposite() : MainThreadNode(8) {
+        cv::namedWindow("Cell Shade", cv::WINDOW_NORMAL);
+        cv::resizeWindow("Cell Shade", 1280, 720);
+    }
+    ~DisplayComposite() { cv::destroyWindow("Cell Shade"); }
+
+    bool operator()(cv::Mat frame) {
+        cv::imshow("Cell Shade", frame);
+        int key = cv::waitKey(1);
+        if (key == 'q' || key == 27) return false;
+        try { return cv::getWindowProperty("Cell Shade", cv::WND_PROP_VISIBLE) >= 1; }
+        catch (const cv::Exception&) { return false; }
+    }
+};
+
+class DisplayEdges : public kpn::MainThreadNode<DisplayEdges,
+                                                kpn::in<"edges">,
+                                                cv::Mat> {
+public:
+    static constexpr std::string_view label()    { return "display_edges"; }
+    static constexpr std::size_t      unique_tag = 1;
+
+    DisplayEdges() : MainThreadNode(8) {
+        cv::namedWindow("Edge Mask", cv::WINDOW_NORMAL);
+        cv::resizeWindow("Edge Mask", 640, 360);
+    }
+    ~DisplayEdges() { cv::destroyWindow("Edge Mask"); }
+
+    bool operator()(cv::Mat mask) {
+        cv::Mat bgr;
+        cv::cvtColor(mask, bgr, cv::COLOR_GRAY2BGR);
+        cv::imshow("Edge Mask", bgr);
+        cv::waitKey(1);
+        try { return cv::getWindowProperty("Edge Mask", cv::WND_PROP_VISIBLE) >= 1; }
+        catch (const cv::Exception&) { return false; }
+    }
+};
+
+// ─────────────────────────────────────────────────────────────────────────────
+
+int main() {
+    using namespace kpn;
+
+    // Nodes — all labelled for web debug UI
+    auto src       = make_node<capture,    "capture">(8);
+    auto gray_node = make_node<to_gray,    "to_gray">(8);
+    auto edge_node = make_node<edges_fn,   "edges"  >(8);
+    auto quant     = make_node<quantise,   "quant"  >(8);
+    auto comp      = make_node<composite,  "comp"   >(8);
+
+    // DisplayNodes live on the main thread — registered as sinks
+    DisplayComposite disp_comp;
+    DisplayEdges     disp_edges;
+
+    // make_network() detects that edge_node.output<0>() feeds two consumers
+    // (comp and disp_edges) and inserts FanoutNode<cv::Mat, 2> automatically.
+    auto net = make_network(
+        edge(src.output<0>(),       quant.input<0>()),       // colour → quant
+        edge(src.output<1>(),       gray_node.input<0>()),   // grey   → to_gray
+        edge(gray_node.output<0>(), edge_node.input<0>()),   // gray   → edges
+        edge(edge_node.output<0>(), comp.input<0>()),        // edges  → comp  (fan-out src)
+        edge(edge_node.output<0>(), disp_edges.input<0>()), // edges  → display_edges (auto fanout)
+        edge(quant.output<0>(),     comp.input<1>()),        // quantised → comp
+        edge(comp.output<0>(),      disp_comp.input<0>())    // result → display_composite
+    );
+
+    std::cout << "Cell-shading pipeline (static) running. Press 'q' to stop.\n";
+
+    net.start();
+
+    // Main thread drives both display nodes — step() returns false when
+    // operator() returns false (q pressed or window closed).
+    while (disp_comp.step() && disp_edges.step())
+        cv::waitKey(8);
+
+    net.stop();
+    return 0;
+}

examples/CMakeLists.txt

@@ -11,6 +11,8 @@ kpn_example(03_multi_output)
 kpn_example(04_storage_policy)
 kpn_example(05_error_handling)
 kpn_example(06_watchdog)
+kpn_example(10_static_hello_pipeline)
+kpn_example(11_static_fanout)
 if(KPN_WEB_DEBUG)
     kpn_target_enable_web_debug(06_watchdog)
 endif()
@@ -24,6 +26,9 @@ find_package(OpenCV QUIET COMPONENTS core imgproc highgui videoio)
 if(OpenCV_FOUND)
     add_executable(09_opencv_cellshade 09_opencv_cellshade/main.cpp)
     target_link_libraries(09_opencv_cellshade PRIVATE kpn ${OpenCV_LIBS})
+
+    add_executable(12_static_cellshade 12_static_cellshade/main.cpp)
+    target_link_libraries(12_static_cellshade PRIVATE kpn ${OpenCV_LIBS})
     if(KPN_WEB_DEBUG)
         kpn_target_enable_web_debug(09_opencv_cellshade)
     endif()

include/kpn/fanout.hpp

@@ -0,0 +1,175 @@
+#pragma once
+#include "channel.hpp"
+#include "diagnostics.hpp"
+#include "node.hpp"
+#include "port.hpp"
+#include "traits.hpp"
+
+#include <array>
+#include <atomic>
+#include <iostream>
+#include <memory>
+#include <thread>
+#include <tuple>
+#include <utility>
+
+namespace kpn {
+
+namespace detail {
+
+// Produces std::tuple<T, T, ..., T> with N repetitions — used so that
+// Network::connect can do its normal return_tuple type-check against FanoutNode.
+template<typename T, std::size_t N, typename Seq = std::make_index_sequence<N>>
+struct repeat_tuple;
+
+template<typename T, std::size_t N, std::size_t... Is>
+struct repeat_tuple<T, N, std::index_sequence<Is...>> {
+    template<std::size_t> using always_T = T;
+    using type = std::tuple<always_T<Is>...>;
+};
+
+template<typename T, std::size_t N>
+using repeat_tuple_t = typename repeat_tuple<T, N>::type;
+
+} // namespace detail
+
+// ── FanoutNode ────────────────────────────────────────────────────────────────
+//
+// Reads one item from its single input channel and pushes a copy to each of
+// N output channels. All N downstream nodes receive every item.
+//
+// Usage:
+//   auto fan = make_fanout<Image, 2>(/*capacity=*/8);
+//   net.connect("src",  src.output<0>(),   "fan",   fan.input<0>())
+//      .connect("fan",  fan.output<0>(),   "nodeA", nodeA.input<0>())
+//      .connect("fan",  fan.output<1>(),   "nodeB", nodeB.input<0>())
+
+template<typename T, std::size_t N>
+class FanoutNode : public INode {
+public:
+    using args_tuple   = std::tuple<T>;
+    using return_tuple = detail::repeat_tuple_t<T, N>;
+    using return_raw   = return_tuple;
+
+    static constexpr std::size_t input_count  = 1;
+    static constexpr std::size_t output_count = N;
+
+    explicit FanoutNode(std::size_t fifo_capacity = 5)
+        : fifo_capacity_(fifo_capacity)
+    {
+        input_ch_ = std::make_shared<Channel<T>>(fifo_capacity);
+    }
+
+    ~FanoutNode() override { stop(); }
+
+    // ── INode ─────────────────────────────────────────────────────────────────
+
+    void start() override {
+        input_ch_->enable();
+        stop_flag_.store(false, std::memory_order_relaxed);
+        thread_ = std::jthread([this](std::stop_token) { run_loop(); });
+    }
+
+    void stop() override {
+        stop_flag_.store(true, std::memory_order_relaxed);
+        input_ch_->disable();
+        if (thread_.joinable()) thread_.request_stop(), thread_.join();
+    }
+
+    bool running() const override {
+        return thread_.joinable() && !stop_flag_.load(std::memory_order_relaxed);
+    }
+
+    void set_name(std::string name) override { name_ = std::move(name); }
+
+    const NodeStats& stats() const override { return stats_; }
+
+    NodeSnapshot node_snapshot(const std::string& name, double elapsed_s) const override {
+        uint64_t frames   = stats_.frames_processed.load(std::memory_order_relaxed);
+        double exec_ms    = stats_.ema_exec_us.load(std::memory_order_relaxed) / 1000.0;
+        double blocked_ms = stats_.total_blocked_us.load(std::memory_order_relaxed) / 1000.0;
+        double total_ms   = exec_ms + blocked_ms;
+        return {name, frames, exec_ms,
+                stats_.max_exec_us.load(std::memory_order_relaxed) / 1000.0,
+                blocked_ms,
+                elapsed_s > 0 ? frames / elapsed_s : 0.0,
+                stats_.total_cpu_us.load(std::memory_order_relaxed) / 1000.0,
+                total_ms > 0 ? 100.0 * exec_ms / total_ms : 0.0};
+    }
+
+    // ── Port access ───────────────────────────────────────────────────────────
+
+    template<std::size_t I = 0>
+    InputPort<FanoutNode, I> input() {
+        static_assert(I == 0, "FanoutNode has exactly one input");
+        return {*this};
+    }
+
+    template<std::size_t I>
+    OutputPort<FanoutNode, I> output() {
+        static_assert(I < N, "FanoutNode output index out of range");
+        return {*this};
+    }
+
+    // ── Internal channel accessors (called by Network::connect) ───────────────
+
+    template<std::size_t I>
+    Channel<T>& input_channel() {
+        static_assert(I == 0);
+        return *input_ch_;
+    }
+
+    template<std::size_t I>
+    void set_input_channel(std::shared_ptr<Channel<T>> ch) {
+        static_assert(I == 0);
+        input_ch_ = std::move(ch);
+    }
+
+    template<std::size_t I>
+    void set_output_channel(Channel<T>* ch) {
+        static_assert(I < N);
+        out_channels_[I] = ch;
+    }
+
+private:
+    void run_loop() {
+        while (!stop_flag_.load(std::memory_order_relaxed)) {
+            try {
+                auto t0  = clock_t::now();
+                T    val = input_ch_->pop();
+                auto t1  = clock_t::now();
+                auto cpu0 = NodeStats::cpu_now();
+
+                for (std::size_t i = 0; i < N; ++i) {
+                    if (out_channels_[i])
+                        out_channels_[i]->push(val);
+                }
+
+                auto cpu1 = NodeStats::cpu_now();
+                auto t2   = clock_t::now();
+                stats_.record_exec(duration_t(t2 - t1), duration_t(t1 - t0), cpu0, cpu1);
+            } catch (const ChannelClosedError&) {
+                break;
+            } catch (const ChannelOverflowError& e) {
+                std::cerr << "[kpn] fanout overflow: " << e.what() << "\n";
+            }
+        }
+    }
+
+    std::string                 name_;
+    std::size_t                 fifo_capacity_;
+    std::shared_ptr<Channel<T>> input_ch_;
+    std::array<Channel<T>*, N>  out_channels_{};
+    std::atomic<bool>           stop_flag_{false};
+    std::jthread                thread_;
+    NodeStats                   stats_;
+};
+
+// ── Factory ───────────────────────────────────────────────────────────────────
+
+template<typename T, std::size_t N>
+FanoutNode<T, N> make_fanout(std::size_t fifo_capacity = 5) {
+    return FanoutNode<T, N>(fifo_capacity);
+}
+
+} // namespace kpn

include/kpn/kpn.hpp

@@ -6,5 +6,7 @@
 #include "channel.hpp"
 #include "port.hpp"
 #include "node.hpp"
+#include "fanout.hpp"
+#include "static_network.hpp"
 #include "main_thread_node.hpp"
 #include "network.hpp"

include/kpn/network.hpp

@@ -12,6 +12,7 @@
 #include <iomanip>
 #include <iostream>
 #include <map>
+#include <set>
 #include <sstream>
 #include <stdexcept>
 #include <string>
@@ -87,6 +88,13 @@ public:
         if (!src) throw NetworkBuildError("node '" + src_name + "' type mismatch");
         if (!dst) throw NetworkBuildError("node '" + dst_name + "' type mismatch");
 
+        auto port_key = std::make_pair(src_name, SrcIdx);
+        if (connected_outputs_.count(port_key))
+            throw NetworkBuildError(
+                "connect: output port '" + src_name + "':" + std::to_string(SrcIdx)
+                + " is already connected — use make_fanout<T, N> for fan-out");
+        connected_outputs_.insert(port_key);
+
         auto& in_ch = dst->template input_channel<DstIdx>();
         src->template set_output_channel<SrcIdx>(&in_ch);
 
@@ -321,6 +329,7 @@ private:
     std::vector<std::string>                               topo_;
     std::map<std::string, std::string>                     exposed_inputs_;
     std::map<std::string, std::string>                     exposed_outputs_;
+    std::set<std::pair<std::string, std::size_t>>          connected_outputs_;
     std::vector<std::unique_ptr<IChannelProbe>>            channel_probes_;
     ErrorHandler                                           error_handler_;
     DiagnosticsHandler                                     diag_handler_;

include/kpn/node.hpp

@@ -37,18 +37,27 @@ struct INode {
 //   InputNames  — optional kpn::in<"a","b"> tag type (at most one)
 //   OutputNames — optional kpn::out<"x","y"> tag type (at most one)
 
-template<auto Func, typename InputTag = in<>, typename OutputTag = out<>>
+template<auto Func,
+         typename InputTag  = in<>,
+         typename OutputTag = out<>,
+         fixed_string Label    = "",
+         std::size_t UniqueTag = 0>
 class Node;
 
 // Specialisation that unpacks the in<>/out<> tag packs
-template<auto Func, fixed_string... InNames, fixed_string... OutNames>
+template<auto Func, fixed_string... InNames, fixed_string... OutNames,
-class Node<Func, in<InNames...>, out<OutNames...>> : public INode {
+         fixed_string Label, std::size_t UniqueTag>
+class Node<Func, in<InNames...>, out<OutNames...>, Label, UniqueTag> : public INode {
 public:
     using F            = decltype(Func);
     using args_tuple   = args_t<F>;
     using return_raw   = return_t<F>;
     using return_tuple = normalised_return_t<return_raw>;
 
+    // Identity accessors — used by StaticNetwork for diagnostics and type-level uniqueness
+    static constexpr std::string_view label()     { return Label.view(); }
+    static constexpr std::size_t      unique_tag  = UniqueTag;
+
     static constexpr std::size_t input_count  = arity_v<F>;
     static constexpr std::size_t output_count = std::tuple_size_v<return_tuple>;
 
@@ -289,17 +298,25 @@ private:
 //   MyFunctor obj(...);
 //   auto node = make_node(obj, in<"x">{}, out<"y">{}, capacity);
 
-template<typename Obj, typename InputTag = in<>, typename OutputTag = out<>>
+template<typename Obj,
+         typename InputTag  = in<>,
+         typename OutputTag = out<>,
+         fixed_string Label    = "",
+         std::size_t UniqueTag = 0>
 class ObjectNode;
 
-template<typename Obj, fixed_string... InNames, fixed_string... OutNames>
+template<typename Obj, fixed_string... InNames, fixed_string... OutNames,
-class ObjectNode<Obj, in<InNames...>, out<OutNames...>> : public INode {
+         fixed_string Label, std::size_t UniqueTag>
+class ObjectNode<Obj, in<InNames...>, out<OutNames...>, Label, UniqueTag> : public INode {
 public:
     using F            = decltype(&Obj::operator());
     using args_tuple   = args_t<F>;
     using return_raw   = return_t<F>;
     using return_tuple = normalised_return_t<return_raw>;
 
+    static constexpr std::string_view label()    { return Label.view(); }
+    static constexpr std::size_t      unique_tag = UniqueTag;
+
     static constexpr std::size_t input_count  = arity_v<F>;
     static constexpr std::size_t output_count = std::tuple_size_v<return_tuple>;
 
@@ -534,32 +551,36 @@ auto make_node(Obj& obj, in<InNames...>, out<OutNames...>, std::size_t fifo_capa
 //
 // Usage:
 //   make_node<func>(capacity)
+//   make_node<func, "label">(capacity)
+//   make_node<func, "label", 1>(capacity)          // UniqueTag=1
 //   make_node<func>(in<"a","b">{}, capacity)
-//   make_node<func>(out<"x">{}, capacity)
+//   make_node<func, "label">(in<"a","b">{}, out<"x">{}, capacity)
-//   make_node<func>(in<"a","b">{}, out<"x">{}, capacity)
 
-// No names
+// No port names
-template<auto Func>
+template<auto Func, fixed_string Label = "", std::size_t UniqueTag = 0>
 auto make_node(std::size_t fifo_capacity = 5) {
-    return Node<Func, in<>, out<>>(fifo_capacity);
+    return Node<Func, in<>, out<>, Label, UniqueTag>(fifo_capacity);
 }
 
 // in<> only
-template<auto Func, fixed_string... InNames>
+template<auto Func, fixed_string Label = "", std::size_t UniqueTag = 0,
+         fixed_string... InNames>
 auto make_node(in<InNames...>, std::size_t fifo_capacity = 5) {
-    return Node<Func, in<InNames...>, out<>>(fifo_capacity);
+    return Node<Func, in<InNames...>, out<>, Label, UniqueTag>(fifo_capacity);
 }
 
-// out<> only (no input names)
+// out<> only
-template<auto Func, fixed_string... OutNames>
+template<auto Func, fixed_string Label = "", std::size_t UniqueTag = 0,
+         fixed_string... OutNames>
 auto make_node(out<OutNames...>, std::size_t fifo_capacity = 5) {
-    return Node<Func, in<>, out<OutNames...>>(fifo_capacity);
+    return Node<Func, in<>, out<OutNames...>, Label, UniqueTag>(fifo_capacity);
 }
 
 // in<> and out<>
-template<auto Func, fixed_string... InNames, fixed_string... OutNames>
+template<auto Func, fixed_string Label = "", std::size_t UniqueTag = 0,
+         fixed_string... InNames, fixed_string... OutNames>
 auto make_node(in<InNames...>, out<OutNames...>, std::size_t fifo_capacity = 5) {
-    return Node<Func, in<InNames...>, out<OutNames...>>(fifo_capacity);
+    return Node<Func, in<InNames...>, out<OutNames...>, Label, UniqueTag>(fifo_capacity);
 }
 
 } // namespace kpn

include/kpn/static_network.hpp

@@ -0,0 +1,237 @@
+#pragma once
+#include "channel.hpp"
+#include "diagnostics.hpp"
+#include "fanout.hpp"
+#include "node.hpp"
+#include "port.hpp"
+#include "tmp/fanout_groups.hpp"
+#include "tmp/topo_sort.hpp"
+
+#include <iostream>
+#include <string>
+#include <tuple>
+#include <type_traits>
+#include <utility>
+
+namespace kpn {
+
+// ── Edge descriptor ───────────────────────────────────────────────────────────
+//
+// Carries compile-time type info and runtime references to the two endpoints.
+// Constructed by the edge() factory; consumed by make_network().
+
+template<typename SrcNode, std::size_t SrcIdx,
+         typename DstNode, std::size_t DstIdx>
+struct Edge {
+    using src_node_t = SrcNode;
+    using dst_node_t = DstNode;
+    static constexpr std::size_t src_idx = SrcIdx;
+    static constexpr std::size_t dst_idx = DstIdx;
+
+    SrcNode& src;
+    DstNode& dst;
+};
+
+template<typename SrcNode, std::size_t SrcIdx,
+         typename DstNode, std::size_t DstIdx>
+auto edge(OutputPort<SrcNode, SrcIdx>, InputPort<DstNode, DstIdx> in)
+    -> Edge<SrcNode, SrcIdx, DstNode, DstIdx>;  // deduction only; defined below
+
+template<typename SrcNode, std::size_t SrcIdx,
+         typename DstNode, std::size_t DstIdx>
+Edge<SrcNode, SrcIdx, DstNode, DstIdx>
+edge(OutputPort<SrcNode, SrcIdx> out, InputPort<DstNode, DstIdx> in) {
+    return {out.node, in.node};
+}
+
+// ── StaticNetwork ─────────────────────────────────────────────────────────────
+
+// node_label<NodeT>: returns Label NTTP as string_view if present, else empty.
+template<typename NodeT, typename = void>
+struct node_label_helper {
+    static constexpr std::string_view value = "";
+};
+template<typename NodeT>
+struct node_label_helper<NodeT, std::void_t<decltype(NodeT::label())>> {
+    static constexpr std::string_view value = NodeT::label();
+};
+
+template<typename NodeT>
+inline constexpr std::string_view node_label_v = node_label_helper<NodeT>::value;
+
+// node_display_name<NodeT, UniqueTag>: Label if non-empty, else "node[UniqueTag]"
+// Returned as std::string at runtime (called once at StaticNetwork construction).
+template<typename NodeT>
+std::string node_display_name() {
+    constexpr std::string_view lbl = node_label_v<NodeT>;
+    if constexpr (!lbl.empty()) {
+        return std::string(lbl);
+    } else if constexpr (requires { NodeT::unique_tag; }) {
+        return "node[" + std::to_string(NodeT::unique_tag) + "]";
+    } else {
+        return "node[?]";
+    }
+}
+
+template<typename FanoutStorage, typename TopoNodeList>
+class StaticNetwork : public INode {
+public:
+    // FanoutStorage = std::tuple<FanoutNode<T0,N0>, ...>  (owned, heap-allocated)
+    // TopoNodeList  = tmp::TypeList<NodeA, NodeB, ...> sources-first
+
+    StaticNetwork(std::unique_ptr<FanoutStorage> fanouts,
+                  std::vector<INode*>        user_nodes_topo,
+                  std::vector<INode*>        fanout_ptrs,
+                  std::vector<std::string>   user_node_names)
+        : fanouts_(std::move(fanouts))
+        , user_nodes_topo_(std::move(user_nodes_topo))
+        , fanout_nodes_ptr_(std::move(fanout_ptrs))
+        , user_node_names_(std::move(user_node_names))
+    {}
+
+    ~StaticNetwork() override { stop(); }
+
+    void start() override {
+        stop_flag_ = false;
+        for (auto* n : user_nodes_topo_)  n->start();
+        for (auto* n : fanout_nodes_ptr_) n->start();
+    }
+
+    void stop() override {
+        stop_flag_ = true;
+        for (auto it = fanout_nodes_ptr_.rbegin(); it != fanout_nodes_ptr_.rend(); ++it)
+            (*it)->stop();
+        for (auto it = user_nodes_topo_.rbegin(); it != user_nodes_topo_.rend(); ++it)
+            (*it)->stop();
+    }
+
+    bool running() const override { return !stop_flag_; }
+
+    void set_name(std::string name) override { name_ = std::move(name); }
+    const NodeStats& stats() const override { static NodeStats dummy; return dummy; }
+    NodeSnapshot node_snapshot(const std::string& n, double) const override {
+        return {n, 0, 0, 0, 0, 0, 0, 0};
+    }
+
+    // Print diagnostics using compile-time node labels
+    void print_diagnostics(std::ostream& os = std::cerr) const {
+        os << "\n┌─ KPN++ StaticNetwork diagnostics ─────────────────────────────\n";
+        for (std::size_t i = 0; i < user_nodes_topo_.size(); ++i) {
+            auto snap = user_nodes_topo_[i]->node_snapshot(user_node_names_[i], 0.0);
+            os << "│  " << snap.name
+               << "  frames=" << snap.frames_processed
+               << "  ema=" << snap.ema_exec_ms << "ms\n";
+        }
+        os << "└────────────────────────────────────────────────────────────────\n";
+    }
+
+    FanoutStorage& fanouts_storage() { return *fanouts_; }
+
+private:
+    std::string                       name_;
+    bool                              stop_flag_{false};
+    std::unique_ptr<FanoutStorage>    fanouts_;
+    std::vector<INode*>               user_nodes_topo_;
+    std::vector<INode*>               fanout_nodes_ptr_;
+    std::vector<std::string>          user_node_names_;   // parallel to user_nodes_topo_
+};
+
+// ── make_network ──────────────────────────────────────────────────────────────
+
+template<typename... Edges>
+auto make_network(Edges&&... edges) {
+    // 1. Expand edges — detect fan-outs, splice FanoutNodes
+    using FanoutSto     = tmp::fanout_storage_t<std::decay_t<Edges>...>;
+    using ExpandedEdges = tmp::expanded_edges_t<std::decay_t<Edges>...>;
+
+    // 2. Duplicate-tag check — fires before cycle check for a cleaner error message
+    using UserNodes = typename tmp::all_node_types<tmp::TypeList<std::decay_t<Edges>...>>::type;
+    static_assert(!tmp::has_duplicate_tags_v<std::decay_t<Edges>...>,
+        "make_network: two nodes have the same (Func, UniqueTag) — they are "
+        "indistinguishable as graph vertices. Add a UniqueTag: "
+        "make_node<func, \"label\", 1>(capacity)");
+
+    // 3. Cycle check
+    using Topo = tmp::topo_sort<ExpandedEdges>;
+    constexpr bool has_cycle = Topo::has_cycle;
+    static_assert(!has_cycle,
+        "make_network: graph contains a directed cycle");
+
+    // 4. Construct owned fanout storage on the heap (FanoutNode has jthread — not moveable)
+    auto fanout_storage = std::make_unique<FanoutSto>();
+
+    // 5. Collect unique user node pointers + their display names, in edge-declaration order
+    std::vector<INode*>      user_node_ptrs;
+    std::vector<std::string> user_node_names;
+    auto collect = [&](auto& e) {
+        using SrcT = std::decay_t<decltype(e.src)>;
+        using DstT = std::decay_t<decltype(e.dst)>;
+        auto* s = static_cast<INode*>(&e.src);
+        auto* d = static_cast<INode*>(&e.dst);
+        if (std::find(user_node_ptrs.begin(), user_node_ptrs.end(), s) == user_node_ptrs.end()) {
+            user_node_ptrs.push_back(s);
+            user_node_names.push_back(node_display_name<SrcT>());
+        }
+        if (std::find(user_node_ptrs.begin(), user_node_ptrs.end(), d) == user_node_ptrs.end()) {
+            user_node_ptrs.push_back(d);
+            user_node_names.push_back(node_display_name<DstT>());
+        }
+    };
+    (collect(edges), ...);
+
+    // 5. Wire all expanded SimpleEdges.
+    //    find_node<NodeT>: searches fanout storage then user edge pack, returns NodeT*.
+    //    Uses if constexpr in a fold so mismatched types never reach assignment.
+    auto find_node = [&]<typename NodeT>() -> NodeT* {
+        NodeT* ptr = nullptr;
+        std::apply([&](auto&... fn) {
+            ([&](auto& node) {
+                if constexpr (std::is_same_v<std::decay_t<decltype(node)>, NodeT>)
+                    if (!ptr) ptr = &node;
+            }(fn), ...);
+        }, *fanout_storage);
+        if (!ptr) {
+            ([&](auto& e) {
+                if (!ptr) {
+                    if constexpr (std::is_same_v<std::decay_t<decltype(e.src)>, NodeT>)
+                        ptr = &e.src;
+                    else if constexpr (std::is_same_v<std::decay_t<decltype(e.dst)>, NodeT>)
+                        ptr = &e.dst;
+                }
+            }(edges), ...);
+        }
+        return ptr;
+    };
+
+    auto wire_one = [&]<typename SE>(SE) {
+        using SrcNode = typename SE::src_node_t;
+        using DstNode = typename SE::dst_node_t;
+        constexpr std::size_t SrcIdx = SE::src_idx;
+        constexpr std::size_t DstIdx = SE::dst_idx;
+
+        auto* src = find_node.template operator()<SrcNode>();
+        auto* dst = find_node.template operator()<DstNode>();
+        if (src && dst)
+            src->template set_output_channel<SrcIdx>(
+                &dst->template input_channel<DstIdx>());
+    };
+
+    [&]<typename... SEs>(tmp::TypeList<SEs...>) {
+        (wire_one(SEs{}), ...);
+    }(ExpandedEdges{});
+
+    // 6. Collect fanout node pointers
+    std::vector<INode*> fanout_ptrs;
+    std::apply([&](auto&... fn) {
+        (fanout_ptrs.push_back(static_cast<INode*>(&fn)), ...);
+    }, *fanout_storage);
+
+    // 7. Construct and return the StaticNetwork
+    using Net = StaticNetwork<FanoutSto, typename Topo::topo>;
+    return Net(std::move(fanout_storage),
+               std::move(user_node_ptrs),
+               std::move(fanout_ptrs),
+               std::move(user_node_names));
+}
+
+} // namespace kpn

include/kpn/tmp/fanout_groups.hpp

@@ -0,0 +1,358 @@
+#pragma once
+#include <cstddef>
+#include <tuple>
+#include <type_traits>
+#include <utility>
+
+// Metafunctions that scan a pack of Edge<> types, group edges by source port,
+// auto-insert FanoutNode<T,N> where N>1, and produce an expanded edge list
+// plus a tuple type for the owned fanout nodes.
+//
+// Key types produced:
+//   expanded_edges_t<Edges...>  — type list of SimpleEdge after fanout insertion
+//   fanout_storage_t<Edges...>  — std::tuple<FanoutNode<T0,N0>, ...> to own
+
+namespace kpn::tmp {
+
+// ── Type list ─────────────────────────────────────────────────────────────────
+
+template<typename... Ts>
+struct TypeList {};
+
+template<typename List, typename T>
+struct append;
+template<typename... Ts, typename T>
+struct append<TypeList<Ts...>, T> { using type = TypeList<Ts..., T>; };
+template<typename List, typename T>
+using append_t = typename append<List, T>::type;
+
+template<typename A, typename B>
+struct concat;
+template<typename... As, typename... Bs>
+struct concat<TypeList<As...>, TypeList<Bs...>> { using type = TypeList<As..., Bs...>; };
+template<typename A, typename B>
+using concat_t = typename concat<A, B>::type;
+
+// ── Source port identity (used as compile-time map key) ───────────────────────
+
+// Two edges share a source port when their SrcNode type AND SrcIdx are identical.
+template<typename SrcNode, std::size_t SrcIdx>
+struct SrcPort {};
+
+template<typename Edge>
+using src_port_of = SrcPort<typename Edge::src_node_t, Edge::src_idx>;
+
+// ── Count occurrences of a SrcPort in an edge list ───────────────────────────
+
+template<typename Port, typename EdgeList>
+struct count_port;
+
+template<typename Port>
+struct count_port<Port, TypeList<>> : std::integral_constant<std::size_t, 0> {};
+
+template<typename Port, typename Head, typename... Tail>
+struct count_port<Port, TypeList<Head, Tail...>>
+    : std::integral_constant<std::size_t,
+          (std::is_same_v<Port, src_port_of<Head>> ? 1 : 0)
+          + count_port<Port, TypeList<Tail...>>::value> {};
+
+// ── Collect all destination (DstNode&, DstIdx) for a given SrcPort ────────────
+
+// A destination descriptor — just type tags, no references (references go in
+// the runtime SimpleEdge structs produced after wiring).
+template<typename DstNode, std::size_t DstIdx>
+struct DstDesc {};
+
+template<typename Port, typename EdgeList>
+struct collect_dsts;
+
+template<typename Port>
+struct collect_dsts<Port, TypeList<>> { using type = TypeList<>; };
+
+template<typename Port, typename Head, typename... Tail>
+struct collect_dsts<Port, TypeList<Head, Tail...>> {
+    using rest = typename collect_dsts<Port, TypeList<Tail...>>::type;
+    using type = std::conditional_t<
+        std::is_same_v<Port, src_port_of<Head>>,
+        append_t<rest, DstDesc<typename Head::dst_node_t, Head::dst_idx>>,
+        rest>;
+};
+
+// ── SimpleEdge: a resolved edge after fanout expansion ────────────────────────
+//
+// At runtime, StaticNetwork wires edges using SimpleEdge descriptors.
+// Each SimpleEdge is just a pair of (SrcNode&, SrcIdx, DstNode&, DstIdx) stored
+// as a type — the actual references come from the node tuple at wire time.
+
+template<typename SrcNode, std::size_t SrcIdx,
+         typename DstNode, std::size_t DstIdx>
+struct SimpleEdge {
+    using src_node_t = SrcNode;
+    using dst_node_t = DstNode;
+    static constexpr std::size_t src_idx = SrcIdx;
+    static constexpr std::size_t dst_idx = DstIdx;
+};
+
+// ── FanoutPlaceholder: a fanout node that will be owned by StaticNetwork ───────
+
+template<typename T, std::size_t N, std::size_t FanoutId>
+struct FanoutPlaceholder {
+    using value_type = T;
+    static constexpr std::size_t fan_n  = N;
+    static constexpr std::size_t fan_id = FanoutId;
+};
+
+// ── expand_edge: for one original edge, produce the replacement SimpleEdge(s) ──
+//
+// If the source port has N>1 consumers: the edge from src→fanout and the
+// fanout→dst edges are synthesised elsewhere (see expand_all). Here we only
+// need to emit the fanout→dst edge for this particular destination.
+//
+// For N==1 edges we emit the edge unchanged.
+//
+// This is called after the fanout node type has already been determined.
+
+} // namespace kpn::tmp
+
+namespace kpn {
+// Forward declaration — FanoutNode is defined in fanout.hpp.
+template<typename T, std::size_t N> class FanoutNode;
+} // namespace kpn
+
+namespace kpn::tmp {
+
+// ── Master expansion: iterate edges, build expanded list + storage tuple ───────
+//
+// Strategy:
+//  1. First pass: for each unique SrcPort with N>1, record a FanoutPlaceholder.
+//  2. Second pass: rewrite each edge.
+//     - N==1 edges become a single SimpleEdge unchanged.
+//     - N>1 edges: on first encounter emit src→fanout SimpleEdge + N fanout→dst
+//       SimpleEdges; on subsequent encounters for the same src port emit nothing
+//       (already handled).
+//
+// To implement "first encounter" tracking we carry a list of already-processed
+// SrcPorts through the fold.
+
+template<typename ProcessedPorts, std::size_t NextFanoutId,
+         typename FanoutList,     // TypeList<FanoutPlaceholder<...>>
+         typename EdgeList,       // TypeList<SimpleEdge<...>> — accumulated output
+         typename RemainingEdges> // TypeList<original edges> still to process
+struct expand_impl;
+
+// Base case — no more edges
+template<typename ProcessedPorts, std::size_t NextFanoutId,
+         typename FanoutList, typename EdgeList>
+struct expand_impl<ProcessedPorts, NextFanoutId, FanoutList, EdgeList, TypeList<>> {
+    using fanouts   = FanoutList;
+    using edges     = EdgeList;
+};
+
+// Helper: is Port in ProcessedPorts?
+template<typename Port, typename Processed>
+struct already_processed : std::false_type {};
+template<typename Port, typename Head, typename... Tail>
+struct already_processed<Port, TypeList<Head, Tail...>>
+    : std::conditional_t<std::is_same_v<Port, Head>,
+          std::true_type,
+          already_processed<Port, TypeList<Tail...>>> {};
+
+// Helper: given DstDesc list + FanoutPlaceholder id, produce SimpleEdge list
+// FanoutNode<T,N>::output<I> → DstNode::input<DstIdx>
+template<std::size_t FanoutId, typename T, std::size_t N,
+         typename DstDescList, std::size_t I = 0>
+struct fanout_to_dst_edges;
+
+template<std::size_t FanoutId, typename T, std::size_t N, std::size_t I>
+struct fanout_to_dst_edges<FanoutId, T, N, TypeList<>, I> {
+    using type = TypeList<>;
+};
+
+template<std::size_t FanoutId, typename T, std::size_t N,
+         typename DstNodeT, std::size_t DstI, typename... DstTail, std::size_t I>
+struct fanout_to_dst_edges<FanoutId, T, N,
+                           TypeList<DstDesc<DstNodeT, DstI>, DstTail...>, I> {
+    using head_edge = SimpleEdge<kpn::FanoutNode<T, N>, I, DstNodeT, DstI>;
+    using rest = typename fanout_to_dst_edges<FanoutId, T, N,
+                                              TypeList<DstTail...>, I+1>::type;
+    using type = append_t<rest, head_edge>;
+};
+
+// Recursive case — process head edge
+template<typename ProcessedPorts, std::size_t NextFanoutId,
+         typename FanoutList, typename EdgeList,
+         typename Head, typename... Tail>
+struct expand_impl<ProcessedPorts, NextFanoutId, FanoutList, EdgeList,
+                   TypeList<Head, Tail...>> {
+
+    using AllEdges = TypeList<Head, Tail...>;
+    using Port     = src_port_of<Head>;
+    using SrcNode  = typename Head::src_node_t;
+    using T        = std::tuple_element_t<Head::src_idx,
+                         typename SrcNode::return_tuple>;
+    static constexpr std::size_t N = count_port<Port, AllEdges>::value
+                                     + count_port<Port, EdgeList>::value
+                                     // recount against full original list approximation:
+                                     // simpler: recount in remaining + already done
+                                     ;
+
+    // Recount properly against the complete original edge list is not possible here
+    // without passing it along. Instead we pre-compute N before entering the fold.
+    // See expand_all below which pre-computes per-port counts.
+    //
+    // This struct is not used directly — expand_all drives the logic with pre-computed N.
+};
+
+// ── expand_all: top-level entry point ─────────────────────────────────────────
+//
+// Pre-computes per-source-port counts, then runs a fold that processes edges
+// one by one.
+
+// Port count map entry
+template<typename Port, std::size_t Count>
+struct PortCount {};
+
+// Build port count list from full edge list
+template<typename AllEdges, typename UniquePortsSeen>
+struct build_port_counts;
+
+template<typename AllEdges>
+struct build_port_counts<AllEdges, TypeList<>> {
+    using type = TypeList<>;
+};
+
+template<typename AllEdges, typename HeadPort, typename... TailPorts>
+struct build_port_counts<AllEdges, TypeList<HeadPort, TailPorts...>> {
+    static constexpr std::size_t cnt = count_port<HeadPort, AllEdges>::value;
+    using rest = typename build_port_counts<AllEdges, TypeList<TailPorts...>>::type;
+    using type = append_t<rest, PortCount<HeadPort, cnt>>;
+};
+
+// Collect unique source ports from edge list
+template<typename EdgeList, typename SeenSoFar = TypeList<>>
+struct unique_src_ports;
+
+template<typename SeenSoFar>
+struct unique_src_ports<TypeList<>, SeenSoFar> { using type = SeenSoFar; };
+
+template<typename Head, typename... Tail, typename SeenSoFar>
+struct unique_src_ports<TypeList<Head, Tail...>, SeenSoFar> {
+    using Port = src_port_of<Head>;
+    using next_seen = std::conditional_t<
+        already_processed<Port, SeenSoFar>::value,
+        SeenSoFar,
+        append_t<SeenSoFar, Port>>;
+    using type = typename unique_src_ports<TypeList<Tail...>, next_seen>::type;
+};
+
+// Look up count for a port
+template<typename Port, typename CountList>
+struct lookup_count : std::integral_constant<std::size_t, 1> {};
+template<typename Port, std::size_t N, typename... Rest>
+struct lookup_count<Port, TypeList<PortCount<Port, N>, Rest...>>
+    : std::integral_constant<std::size_t, N> {};
+template<typename Port, typename Head, typename... Rest>
+struct lookup_count<Port, TypeList<Head, Rest...>>
+    : lookup_count<Port, TypeList<Rest...>> {};
+
+// Fold state for the wiring pass
+template<typename ProcessedPorts, std::size_t NextFanoutId,
+         typename FanoutList, typename EdgeList>
+struct FoldState {
+    using processed = ProcessedPorts;
+    static constexpr std::size_t next_id = NextFanoutId;
+    using fanouts = FanoutList;
+    using edges   = EdgeList;
+};
+
+// Process one edge given pre-computed port counts
+template<typename State, typename Edge, typename CountList, typename AllEdges>
+struct process_edge {
+    using Port    = src_port_of<Edge>;
+    using SrcNode = typename Edge::src_node_t;
+    using T       = std::tuple_element_t<Edge::src_idx, typename SrcNode::return_tuple>;
+    static constexpr std::size_t N = lookup_count<Port, CountList>::value;
+
+    // N==1: pass through unchanged
+    using passthrough_edges = append_t<typename State::edges,
+        SimpleEdge<SrcNode, Edge::src_idx,
+                   typename Edge::dst_node_t, Edge::dst_idx>>;
+
+    // N>1, first encounter: emit src→fanout + all fanout→dst edges
+    using DstDescs = typename collect_dsts<Port, AllEdges>::type;
+    static constexpr std::size_t fid = State::next_id;
+    using fanout_type = FanoutPlaceholder<T, N, fid>;
+    using src_to_fan  = SimpleEdge<SrcNode, Edge::src_idx, FanoutNode<T, N>, 0>;
+    using fan_to_dsts = typename fanout_to_dst_edges<fid, T, N, DstDescs>::type;
+    using fanout_edges = concat_t<append_t<typename State::edges, src_to_fan>, fan_to_dsts>;
+    using new_fanouts  = append_t<typename State::fanouts, fanout_type>;
+
+    static constexpr bool seen = already_processed<Port, typename State::processed>::value;
+
+    using type = std::conditional_t<
+        (N == 1),
+        FoldState<typename State::processed, State::next_id,
+                  typename State::fanouts, passthrough_edges>,
+        std::conditional_t<
+            !seen,
+            FoldState<append_t<typename State::processed, Port>,
+                      State::next_id + 1,
+                      new_fanouts, fanout_edges>,
+            // already processed — skip (fanout edges already emitted)
+            State>>;
+};
+
+// Fold over all edges
+template<typename State, typename EdgeList, typename CountList, typename AllEdges>
+struct fold_edges;
+
+template<typename State, typename CountList, typename AllEdges>
+struct fold_edges<State, TypeList<>, CountList, AllEdges> { using type = State; };
+
+template<typename State, typename Head, typename... Tail,
+         typename CountList, typename AllEdges>
+struct fold_edges<State, TypeList<Head, Tail...>, CountList, AllEdges> {
+    using next  = typename process_edge<State, Head, CountList, AllEdges>::type;
+    using type  = typename fold_edges<next, TypeList<Tail...>, CountList, AllEdges>::type;
+};
+
+// ── Public interface ───────────────────────────────────────────────────────────
+
+template<typename... Edges>
+struct expand_all {
+    using AllEdges     = TypeList<Edges...>;
+    using UniquePorts  = typename unique_src_ports<AllEdges>::type;
+    using CountList    = typename build_port_counts<AllEdges, UniquePorts>::type;
+    using InitState    = FoldState<TypeList<>, 0, TypeList<>, TypeList<>>;
+    using FinalState   = typename fold_edges<InitState, AllEdges, CountList, AllEdges>::type;
+
+    using fanout_placeholders = typename FinalState::fanouts;  // TypeList<FanoutPlaceholder<...>>
+    using expanded_edges      = typename FinalState::edges;    // TypeList<SimpleEdge<...>>
+};
+
+// Convert TypeList<FanoutPlaceholder<T,N,Id>...> to std::tuple<FanoutNode<T,N>...>
+template<typename PlaceholderList>
+struct to_fanout_tuple;
+
+template<>
+struct to_fanout_tuple<TypeList<>> { using type = std::tuple<>; };
+
+template<typename T, std::size_t N, std::size_t Id, typename... Rest>
+struct to_fanout_tuple<TypeList<FanoutPlaceholder<T, N, Id>, Rest...>> {
+    using rest = typename to_fanout_tuple<TypeList<Rest...>>::type;
+    // prepend FanoutNode<T,N>
+    template<typename Tuple> struct prepend;
+    template<typename... Ts> struct prepend<std::tuple<Ts...>> {
+        using type = std::tuple<kpn::FanoutNode<T, N>, Ts...>;
+    };
+    using type = typename prepend<rest>::type;
+};
+
+template<typename... Edges>
+using fanout_storage_t = typename to_fanout_tuple<
+    typename expand_all<Edges...>::fanout_placeholders>::type;
+
+template<typename... Edges>
+using expanded_edges_t = typename expand_all<Edges...>::expanded_edges;
+
+} // namespace kpn::tmp

include/kpn/tmp/repeat_tuple.hpp

@@ -0,0 +1,23 @@
+#pragma once
+#include <cstddef>
+#include <tuple>
+#include <utility>
+
+namespace kpn::tmp {
+
+// repeat_tuple_t<T, N> — std::tuple<T, T, ..., T> with N elements.
+// Used by FanoutNode and fanout_groups to express N homogeneous outputs.
+
+template<typename T, std::size_t N, typename Seq = std::make_index_sequence<N>>
+struct repeat_tuple;
+
+template<typename T, std::size_t N, std::size_t... Is>
+struct repeat_tuple<T, N, std::index_sequence<Is...>> {
+    template<std::size_t> using always_T = T;
+    using type = std::tuple<always_T<Is>...>;
+};
+
+template<typename T, std::size_t N>
+using repeat_tuple_t = typename repeat_tuple<T, N>::type;
+
+} // namespace kpn::tmp

include/kpn/tmp/topo_sort.hpp

@@ -0,0 +1,223 @@
+#pragma once
+#include "fanout_groups.hpp"
+#include <cstddef>
+#include <tuple>
+#include <type_traits>
+
+// Compile-time topological sort + cycle detection over a TypeList<SimpleEdge<...>>.
+//
+// Nodes are identified by type (not by index), since two different node types
+// are always distinct vertices. The sort produces a TypeList of node types in
+// topological order (sources first).
+//
+// cycle detection: if DFS revisits a node currently on the stack → static_assert.
+
+namespace kpn::tmp {
+
+// ── Collect all unique node types from an expanded edge list ──────────────────
+
+template<typename EdgeList, typename Seen = TypeList<>>
+struct all_node_types;
+
+template<typename Seen>
+struct all_node_types<TypeList<>, Seen> { using type = Seen; };
+
+template<typename Head, typename... Tail, typename Seen>
+struct all_node_types<TypeList<Head, Tail...>, Seen> {
+    using S1 = std::conditional_t<
+        already_processed<typename Head::src_node_t, Seen>::value,
+        Seen, append_t<Seen, typename Head::src_node_t>>;
+    using S2 = std::conditional_t<
+        already_processed<typename Head::dst_node_t, S1>::value,
+        S1, append_t<S1, typename Head::dst_node_t>>;
+    using type = typename all_node_types<TypeList<Tail...>, S2>::type;
+};
+
+// ── Collect successors (dst node types) of a given src node type ──────────────
+
+template<typename NodeType, typename EdgeList>
+struct successors;
+
+template<typename NodeType>
+struct successors<NodeType, TypeList<>> { using type = TypeList<>; };
+
+template<typename NodeType, typename Head, typename... Tail>
+struct successors<NodeType, TypeList<Head, Tail...>> {
+    using rest = typename successors<NodeType, TypeList<Tail...>>::type;
+    using type = std::conditional_t<
+        std::is_same_v<typename Head::src_node_t, NodeType>,
+        std::conditional_t<
+            already_processed<typename Head::dst_node_t, rest>::value,
+            rest,
+            append_t<rest, typename Head::dst_node_t>>,
+        rest>;
+};
+
+// ── DFS state ─────────────────────────────────────────────────────────────────
+
+// Color: 0=white(unseen), 1=grey(on stack), 2=black(done)
+// We track grey nodes as a TypeList to detect back-edges.
+
+template<typename Node, typename GreyList>
+struct is_grey : already_processed<Node, GreyList> {};
+
+// DFS result: has_cycle flag + topo order (black nodes appended post-visit)
+template<bool Cycle, typename TopoList>
+struct DfsResult { static constexpr bool has_cycle = Cycle; using topo = TopoList; };
+
+template<typename Node, typename EdgeList,
+         typename GreyList, typename BlackList, typename TopoList>
+struct dfs_node;
+
+// Iterate successors
+template<typename SuccList, typename EdgeList,
+         typename GreyList, typename BlackList, typename TopoList, bool CycleSoFar>
+struct dfs_successors;
+
+template<typename EdgeList, typename GreyList, typename BlackList,
+         typename TopoList, bool CycleSoFar>
+struct dfs_successors<TypeList<>, EdgeList, GreyList, BlackList, TopoList, CycleSoFar> {
+    static constexpr bool has_cycle = CycleSoFar;
+    using black  = BlackList;
+    using topo   = TopoList;
+};
+
+template<typename Head, typename... Tail, typename EdgeList,
+         typename GreyList, typename BlackList, typename TopoList, bool CycleSoFar>
+struct dfs_successors<TypeList<Head, Tail...>, EdgeList,
+                      GreyList, BlackList, TopoList, CycleSoFar> {
+    // Visit Head
+    using visit = dfs_node<Head, EdgeList, GreyList, BlackList, TopoList>;
+    static constexpr bool cycle1 = CycleSoFar || visit::has_cycle;
+    // Continue with remaining successors using updated black/topo
+    using rest = dfs_successors<TypeList<Tail...>, EdgeList,
+                                GreyList, typename visit::black,
+                                typename visit::topo, cycle1>;
+    static constexpr bool has_cycle = rest::has_cycle;
+    using black = typename rest::black;
+    using topo  = typename rest::topo;
+};
+
+template<typename Node, typename EdgeList,
+         typename GreyList, typename BlackList, typename TopoList>
+struct dfs_node {
+    // Already black — skip
+    static constexpr bool already_done = already_processed<Node, BlackList>::value;
+    // On stack — cycle
+    static constexpr bool on_stack     = is_grey<Node, GreyList>::value;
+
+    // Visit successors (only if not already done / on stack)
+    using succs = typename successors<Node, EdgeList>::type;
+    using new_grey = append_t<GreyList, Node>;
+
+    using visit_succs = dfs_successors<succs, EdgeList, new_grey, BlackList, TopoList,
+                                       on_stack>;
+
+    static constexpr bool has_cycle = already_done ? false :
+                                      on_stack     ? true  :
+                                      visit_succs::has_cycle;
+
+    // Append node to topo after all successors (post-order = reverse topo)
+    using topo_after = std::conditional_t<
+        already_done || on_stack,
+        TopoList,
+        append_t<typename visit_succs::topo, Node>>;
+
+    using black = std::conditional_t<
+        already_done || on_stack,
+        BlackList,
+        append_t<typename visit_succs::black, Node>>;
+
+    using topo = topo_after;
+};
+
+// ── Top-level DFS over all nodes ──────────────────────────────────────────────
+
+template<typename NodeList, typename EdgeList,
+         typename BlackList, typename TopoList, bool CycleSoFar>
+struct dfs_all;
+
+template<typename EdgeList, typename BlackList, typename TopoList, bool CycleSoFar>
+struct dfs_all<TypeList<>, EdgeList, BlackList, TopoList, CycleSoFar> {
+    static constexpr bool has_cycle = CycleSoFar;
+    using topo  = TopoList;
+};
+
+template<typename Head, typename... Tail, typename EdgeList,
+         typename BlackList, typename TopoList, bool CycleSoFar>
+struct dfs_all<TypeList<Head, Tail...>, EdgeList, BlackList, TopoList, CycleSoFar> {
+    using visit = dfs_node<Head, EdgeList, TypeList<>, BlackList, TopoList>;
+    static constexpr bool cycle1 = CycleSoFar || visit::has_cycle;
+    using rest = dfs_all<TypeList<Tail...>, EdgeList,
+                         typename visit::black, typename visit::topo, cycle1>;
+    static constexpr bool has_cycle = rest::has_cycle;
+    using topo = typename rest::topo;
+};
+
+// ── Public interface ───────────────────────────────────────────────────────────
+
+// topo_sort<EdgeList>:
+//   ::topo      — TypeList of node types, sources first (start order)
+//   ::has_cycle — true if a cycle was detected
+template<typename EdgeList>
+struct topo_sort {
+    using Nodes = typename all_node_types<EdgeList>::type;
+    using result = dfs_all<Nodes, EdgeList, TypeList<>, TypeList<>, false>;
+    // DFS post-order gives reverse topo; reverse the list for sources-first order.
+    // Reversing a TypeList:
+    template<typename List, typename Acc = TypeList<>>
+    struct reverse_list;
+    template<typename Acc>
+    struct reverse_list<TypeList<>, Acc> { using type = Acc; };
+    template<typename H, typename... T, typename Acc>
+    struct reverse_list<TypeList<H, T...>, Acc>
+        : reverse_list<TypeList<T...>, append_t<Acc, H>> {};  // wrong direction intentionally
+
+    // Post-order appends children before parent, so the list is already
+    // reverse-topo (sinks first). Reverse it to get sources first.
+    template<typename List, typename Acc = TypeList<>>
+    struct rev;
+    template<typename Acc>
+    struct rev<TypeList<>, Acc> { using type = Acc; };
+    template<typename H, typename... T, typename Acc>
+    struct rev<TypeList<H, T...>, Acc> : rev<TypeList<T...>, TypeList<H, Acc>> {
+        // prepend H to Acc: TypeList<H, Acc...>
+    };
+    // Simpler: just reverse by prepending
+    template<typename L, typename A = TypeList<>>
+    struct rev2 { using type = A; };
+    template<typename H, typename... T, typename... As>
+    struct rev2<TypeList<H, T...>, TypeList<As...>>
+        : rev2<TypeList<T...>, TypeList<H, As...>> {};
+
+    static constexpr bool has_cycle = result::has_cycle;
+    using topo = typename rev2<typename result::topo>::type;
+};
+
+// ── Duplicate-tag detection ───────────────────────────────────────────────────
+//
+// Two user nodes share a (Func, UniqueTag) pair iff they have the same type.
+// has_duplicate_tags_v<Edges...> is true if any two user node types are identical.
+
+template<typename T, typename List>
+struct type_in_list : std::false_type {};
+template<typename T, typename H, typename... Tail>
+struct type_in_list<T, TypeList<H, Tail...>>
+    : std::conditional_t<std::is_same_v<T, H>,
+          std::true_type,
+          type_in_list<T, TypeList<Tail...>>> {};
+
+template<typename NodeList>
+struct has_duplicate_node_types : std::false_type {};
+template<typename H, typename... Tail>
+struct has_duplicate_node_types<TypeList<H, Tail...>>
+    : std::conditional_t<type_in_list<H, TypeList<Tail...>>::value,
+          std::true_type,
+          has_duplicate_node_types<TypeList<Tail...>>> {};
+
+template<typename... Edges>
+inline constexpr bool has_duplicate_tags_v =
+    has_duplicate_node_types<
+        typename all_node_types<TypeList<Edges...>>::type>::value;
+
+} // namespace kpn::tmp

tests/CMakeLists.txt

@@ -31,6 +31,7 @@ add_executable(kpn_tests
     test_channel.cpp
     test_node.cpp
     test_network.cpp
+    test_static_network.cpp
 )
 
 target_link_libraries(kpn_tests PRIVATE

tests/test_static_network.cpp

@@ -0,0 +1,209 @@
+#include <catch2/catch_test_macros.hpp>
+#include <kpn/kpn.hpp>
+#include <chrono>
+#include <thread>
+#include <atomic>
+
+using namespace kpn;
+
+static int increment(int x)  { return x + 1; }
+static int multiply2(int x)  { return x * 2; }
+static int multiply3(int x)  { return x * 3; }
+static int add10(int x)      { return x + 10; }
+static int negate_val(int x) { return -x; }
+static int square(int x)     { return x * x; }
+
+// ── Linear pipeline ───────────────────────────────────────────────────────────
+
+TEST_CASE("static_network: linear pipeline produces correct result", "[static_network]") {
+    auto src = make_node<increment>(5);
+    auto dst = make_node<multiply2>(5);
+
+    Channel<int> final_out(5);
+    dst.set_output_channel<0>(&final_out);
+
+    auto net = make_network(
+        edge(src.output<0>(), dst.input<0>())
+    );
+
+    net.start();
+    src.input_channel<0>().push(5);   // 5 → increment → 6 → multiply2 → 12
+    int result = final_out.pop();
+    net.stop();
+
+    REQUIRE(result == 12);
+}
+
+TEST_CASE("static_network: three-node pipeline", "[static_network]") {
+    auto a = make_node<increment>(5);
+    auto b = make_node<multiply2>(5);
+    auto c = make_node<add10>(5);
+
+    Channel<int> out(5);
+    c.set_output_channel<0>(&out);
+
+    auto net = make_network(
+        edge(a.output<0>(), b.input<0>()),
+        edge(b.output<0>(), c.input<0>())
+    );
+
+    net.start();
+    a.input_channel<0>().push(3);   // 3 → +1=4 → *2=8 → +10=18
+    int result = out.pop();
+    net.stop();
+
+    REQUIRE(result == 18);
+}
+
+// ── Auto fan-out ──────────────────────────────────────────────────────────────
+
+TEST_CASE("static_network: auto fanout delivers to both consumers", "[static_network]") {
+    // Use distinct functions so each node has a distinct type in the graph
+    auto src  = make_node<increment>(8);
+    auto dstA = make_node<multiply2>(8);
+    auto dstB = make_node<multiply3>(8);
+
+    Channel<int> outA(8), outB(8);
+    dstA.set_output_channel<0>(&outA);
+    dstB.set_output_channel<0>(&outB);
+
+    // Two edges from the same output port — FanoutNode<int,2> is auto-inserted
+    auto net = make_network(
+        edge(src.output<0>(), dstA.input<0>()),
+        edge(src.output<0>(), dstB.input<0>())
+    );
+
+    net.start();
+    src.input_channel<0>().push(3);   // 3 → +1=4 → *2=8 and *3=12
+
+    int a = outA.pop();
+    int b = outB.pop();
+    net.stop();
+
+    REQUIRE(a == 8);
+    REQUIRE(b == 12);
+}
+
+TEST_CASE("static_network: auto fanout preserves ordering across multiple items", "[static_network]") {
+    auto src  = make_node<increment>(16);
+    auto dstA = make_node<multiply2>(16);
+    auto dstB = make_node<negate_val>(16);
+
+    Channel<int> outA(16), outB(16);
+    dstA.set_output_channel<0>(&outA);
+    dstB.set_output_channel<0>(&outB);
+
+    auto net = make_network(
+        edge(src.output<0>(), dstA.input<0>()),
+        edge(src.output<0>(), dstB.input<0>())
+    );
+
+    net.start();
+    for (int i = 0; i < 5; ++i)
+        src.input_channel<0>().push(i);   // 0..4 → +1 → *2 or negate
+
+    for (int i = 0; i < 5; ++i) {
+        REQUIRE(outA.pop() == (i + 1) * 2);
+        REQUIRE(outB.pop() == -(i + 1));
+    }
+    net.stop();
+}
+
+// ── Stop/start lifecycle ──────────────────────────────────────────────────────
+
+TEST_CASE("static_network: stop disables input channel", "[static_network]") {
+    auto src = make_node<increment>(5);
+    auto dst = make_node<multiply2>(5);
+
+    auto net = make_network(
+        edge(src.output<0>(), dst.input<0>())
+    );
+
+    net.start();
+    net.stop();
+
+    // After stop, input channel disabled — push must not throw
+    src.input_channel<0>().push(99);
+    REQUIRE(src.input_channel<0>().size() == 0);
+}
+
+// ── Compile-time cycle detection ──────────────────────────────────────────────
+// Cycles fire a static_assert in make_network(), so we can only test the
+// no-cycle path at runtime.
+//
+// To manually verify a cycle error: add
+//   auto bad = make_network(edge(a.output<0>(), b.input<0>()),
+//                           edge(b.output<0>(), a.input<0>()));
+// and confirm: "make_network: graph contains a directed cycle"
+//
+// To manually verify a duplicate-tag error: add
+//   auto x = make_node<increment>(5);
+//   auto y = make_node<increment>(5);   // same type as x — UniqueTag=0 for both
+//   auto bad = make_network(edge(x.output<0>(), y.input<0>()));
+// and confirm: "make_network: two nodes have the same (Func, UniqueTag)"
+
+TEST_CASE("static_network: acyclic graph does not trigger static_assert", "[static_network]") {
+    auto a = make_node<increment>(5);
+    auto b = make_node<add10>(5);
+
+    Channel<int> out(5);
+    b.set_output_channel<0>(&out);
+
+    auto net = make_network(edge(a.output<0>(), b.input<0>()));
+
+    net.start();
+    a.input_channel<0>().push(5);
+    REQUIRE(out.pop() == 16);   // 5 → +1=6 → +10=16
+    net.stop();
+}
+
+// ── Label and UniqueTag ───────────────────────────────────────────────────────
+
+TEST_CASE("static_network: same function distinguished by UniqueTag", "[static_network]") {
+    // Two nodes wrapping the same function — only possible with distinct UniqueTag
+    auto a = make_node<increment, "stage1", 0>(8);
+    auto b = make_node<increment, "stage2", 1>(8);  // same func, tag=1 → distinct type
+
+    Channel<int> out(8);
+    b.set_output_channel<0>(&out);
+
+    auto net = make_network(edge(a.output<0>(), b.input<0>()));
+
+    net.start();
+    a.input_channel<0>().push(10);
+    REQUIRE(out.pop() == 12);   // 10 → +1=11 → +1=12
+    net.stop();
+}
+
+TEST_CASE("static_network: label is accessible as static member", "[static_network]") {
+    using MyNode = decltype(make_node<increment, "my_node">(5));
+    REQUIRE(MyNode::label() == "my_node");
+    REQUIRE(MyNode::unique_tag == 0);
+
+    using TaggedNode = decltype(make_node<increment, "tagged", 42>(5));
+    REQUIRE(TaggedNode::label() == "tagged");
+    REQUIRE(TaggedNode::unique_tag == 42);
+}
+
+TEST_CASE("static_network: fanout with labelled same-function consumers", "[static_network]") {
+    auto src  = make_node<increment, "src"         >(8);
+    auto dstA = make_node<increment, "consumer_a", 1>(8);
+    auto dstB = make_node<increment, "consumer_b", 2>(8);
+
+    Channel<int> outA(8), outB(8);
+    dstA.set_output_channel<0>(&outA);
+    dstB.set_output_channel<0>(&outB);
+
+    // Fan-out from src to two increment nodes — only possible because tags differ
+    auto net = make_network(
+        edge(src.output<0>(), dstA.input<0>()),
+        edge(src.output<0>(), dstB.input<0>())
+    );
+
+    net.start();
+    src.input_channel<0>().push(5);   // 5 → +1=6 → both +1=7
+
+    REQUIRE(outA.pop() == 7);
+    REQUIRE(outB.pop() == 7);
+    net.stop();
+}