FLOW: The FLOW Field Guide—and the Questions Still Open

On this page↓

A good first visit does not need to explain every equation. It needs to give you one reliable observation and a way to avoid believing too much from it.

For FLOW, begin by identifying the object in front of you: the public scene is a recorded preview. The small article explorations are teaching examples or selected measured records. The live fluid laboratory remains separate and under review.

A useful first minute

Open the recorded preview. Before pressing Play recording, read the still's direct labels. Identify A and B, their viscosities and the matched model time. The poster belongs to the fixed Keep the Swirl example even if another guide is selected.

Open About this run. Read the explanation of what the selected guide describes and what the fixed recording actually contains. This is where a setting, an observation and a caveat should meet.

Then play the recording if you want motion. It does not start a new solver run. If playback is unavailable or reduced motion is preferred, the real still and transcript remain useful. A fallback should preserve the explanation rather than replace it with a false live claim.

Four questions for the four guides

Make a Flow: distinguish an input from the motion that continues after it. The public guide explains this question; it does not let you apply an arbitrary live force to the recording.

Keep the Swirl: read the controlled viscosity comparison. Check the common starting state and observation time before interpreting the difference.

Shape the Wake: identify the domain and obstacle question. A channel guide does not turn a periodic recording into a validated wake experiment.

Read the Invisible: separate concentration, speed, direction and signed vorticity. Switching an explanation is not equivalent to producing a fresh field measurement.

I identified the recording and its A/B settings.I distinguished dye, speed and signed vorticity.I checked that the compared samples share a time.I separated a measured result from a general proof.

0 / 4 · questions checked

Your reading checklist stays in this page’s memory. Checking a box is not an assessment, human review or scientific qualification.

The checklist is personal and temporary. It does not send a learning score or create a scientific approval. Its purpose is to turn a passive visit into four explicit reading actions.

Preserve the right thing

A captioned image is a record of a particular displayed result. Read its actual experiment title and time rather than assuming it inherits every guide selection. A setup link describes supported initial settings; it does not preserve your gestures or promise a replay of an unrecorded live experiment.

The release manifest identifies the historical recording assets. The article measurement file identifies the old and corrected numerical runs used in this series. They answer different provenance questions. Keep both identities when sharing a comparison.

“Same source” and “same picture” are also different claims. A new renderer can display the same data differently. A new tracer method can produce different data under a similar-looking image. The version tells you which question to ask next; it does not replace the question.

What has been established

The focused current periodic checks characterize named fields, grids, parameter choices and durations on the recorded test hardware. The conservation repair passes the unchanged budget for the two specified compact-vortex cases. Equality, replay, operator and stage checks each add evidence for their own contracts.

That evidence does not complete the general-engine catalog, developed-wake and outlet sensitivity work, physical-phone qualification, sustained thermal/frame-pacing checks, manual accessibility studies or human scientific and fluent-reader sign-offs. A successful deployment is not a substitute for those activities.

The series itself can still be useful now. It teaches the distinction between a measured result and a general claim without requiring you to trust a moving image on instinct.

The question to keep

Return to the first article's beautiful vortex. You can now ask what the color means, whether a curve is a trajectory, what domain the computation lives in, what changed between panels, and which failed test would make you revise your interpretation.

That is a better destination than memorizing the controls. The instrument is useful when it changes the quality of your questions.

Carry this into the next project: what was held fixed, what was measured, and what observation would make the explanation fail?

Return to Inside FLOW, or read the project overview with those questions in mind.