Skip to content

ARCHITECTURE: drop references to the concept draft #21

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ehsun-sh merged 1 commit into from
Oct 8, 2026
Merged

ARCHITECTURE: drop references to the concept draft #21

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 7 additions & 8 deletions docs/ARCHITECTURE.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@

**Status:** design document, now describing a running system. Phases 0 through 5 are implemented
and released as 0.x; what each block still approximates is collected under [§10, Open](#open).
**Revision:** 2 — restructured, corrected, and expanded from the initial concept draft.
**Revision:** 2.

---

Expand Down Expand Up @@ -103,8 +103,8 @@ it rather than rewriting it.

Analysis and visualization are **not** a layer in this stack. They are *consumers* of results —
some run in the engine (eye-histogram binning, spectrum estimation, BER counting), some run in
the browser (rendering). Placing them as a layer between the GUI and the component library, as
the initial draft did, describes no real dependency and would produce a confused codebase.
the browser (rendering). Placing them as a layer between the GUI and the component library
describes no real dependency and would produce a confused codebase.

### 2.2 Layer responsibilities

Expand Down Expand Up @@ -153,8 +153,8 @@ class SimulationContext:
```

Every sampled signal in a run shares `N`, `Fs`, and the same time origin. This is what makes
blocks composable at all, and it is what the initial draft's per-signal `sampling_rate` /
`num_samples` fields quietly allowed to drift apart.
blocks composable at all; per-signal `sampling_rate` / `num_samples` fields would quietly
allow them to drift apart.

**Reproducibility rule:** all stochastic blocks (laser phase noise, ASE, shot noise, thermal
noise) draw from independent generators derived deterministically from `seed` and the block's
Expand Down Expand Up @@ -257,7 +257,7 @@ suite runs in both.

## 4. Execution Engine

Absent from the original draft, and the heart of any dataflow simulator.
The heart of any dataflow simulator.

### 4.1 Execution model

Expand Down Expand Up @@ -441,8 +441,7 @@ headless with no GUI installed.
architecture. A browser-accessible tool removes the largest barrier to trying it.
* **Graph editor:** use an existing canvas library (React Flow / rete.js). The node-graph editor
is a solved problem and not where this project's value lies.
* **Plot data is reduced in the engine, not in the browser.** This is the key point the original
draft missed. An eye diagram is a 2D histogram — the engine bins millions of samples and sends
* **Plot data is reduced in the engine, not in the browser.** This is the key point. An eye diagram is a 2D histogram — the engine bins millions of samples and sends
a small array; the browser never receives raw sample buffers. Same for spectra (send the
computed PSD on a display grid) and BER curves. With this rule, ordinary Canvas/WebGL2
rendering is more than sufficient.
Expand Down
Loading