Distance sounds like it has one meaning. In metrology it has at least three, and the difference between them shows up precisely when the part is not perfectly square to the camera — which is always.
Two different questions
Consider a slot and a hole, and a drawing calling out the distance between them.
"How close do these get?"
This is a shortest distance. It is the minimum over every pair of points on the two features, and its direction is whatever direction happens to be shortest. It is the right question for clearance, interference, wall thickness and minimum material condition — anything where what matters is the tightest place, wherever that is.
"How far apart are they, measured across?"
This is a fixed-direction distance. You have decided the direction in advance — horizontal, vertical, or perpendicular to a stated datum — and you are measuring the separation along it. It is the right question for coordinate dimensions, centre distances referenced to an edge, and anything the drawing dimensions from a datum.
These give the same answer only when the shortest direction happens to coincide with the fixed one. On a real part, imaged by a real operator, it usually does not.
fixed-direction ≥ shortest, always
The shortest distance is a minimum, so any other direction can only be longer. A measurement that is systematically the smallest possible value is not conservative — it is optimistic for clearance and pessimistic for coverage, and either way it is not what the drawing asked for.
The rotation trap
Here is the failure that actually costs money. A part is dimensioned 40.00 mm horizontally between two edges. The operator places it under the camera slightly rotated — say 3°, which looks like nothing on screen.
| Tool | Reads | Why |
|---|---|---|
| Point-to-point | 40.00 mm | It measures along the part, which is the true feature length. |
| Horizontal | about 39.95 mm | It measures the horizontal component, and the part is no longer horizontal. |
Both tools worked correctly. But if the drawing means "the length of this feature", the horizontal tool has introduced an error that depends entirely on how the operator laid the part down — and it will vary from part to part, looking exactly like process variation.
Now invert it. If the dimension is genuinely referenced to a horizontal datum — a coordinate position measured across from a reference edge — then the horizontal tool is right and point-to-point is wrong, because the part being rotated genuinely does change that coordinate.
The question is not which tool is more accurate. It is which one matches the drawing. And if the part's orientation in the frame changes the answer, fixture the part — do not choose a tool that hides it.
Which one does the drawing mean?
- Feature length or width — the extent of the feature itself: point-to-point along it.
- Dimensioned from a datum edge — the direction is set by the datum: horizontal, vertical, or perpendicular to that edge.
- Clearance, gap or minimum wall — shortest distance, because the tightest point is the point that matters.
- Maximum extent or envelope — farthest distance.
- Standoff from an edge or centreline — perpendicular to that line.
When the drawing is ambiguous — and drawings often are — the right move is to ask and then record the decision, not to pick the tool that produces the most comfortable number. Write the chosen convention into the pattern row description so the next person measures it the same way.
Perpendicular, and to what
"Perpendicular distance" needs a reference, and the reference is a choice. Perpendicular to the part edge, to a construction guide, or to the image axes are three different measurements.
This is why construction geometry matters more than it first appears. Drawing an explicit datum line along the reference edge — and then measuring perpendicular to that — makes the convention visible on the image rather than living in the operator's head. Anyone reading the report afterwards can see what the measurement was taken against.
The same applies to arcs. A distance to an arc taken as "shortest" runs along the radial direction to the nearest point. A distance taken along the chord normal is measured in a direction the arc itself defines. On a shallow arc these can differ noticeably, and only one of them is what a radius-referenced callout means.
How Measuret separates them
Measuret does not have one "distance" button with hidden behaviour. Each convention is a separate tool with its own name, so the choice is explicit and visible in the report:
| Tool | Direction |
|---|---|
| Point-Point | Whatever direction the two points define |
| Horizontal | Fixed: horizontal component only |
| Vertical | Fixed: vertical component only |
| Pt-Line | Perpendicular to the chosen line |
| Pt-Arc | Shortest, to the arc |
| Pt-Arc Normal | Along the chord normal the arc defines |
| Closest / Farthest | Minimum / maximum between two whole shapes |
| Perp Closest / Perp Farthest | Minimum / maximum, constrained perpendicular to a reference |
Construction guides — including an offset guide placed at a distance you type — let you build the datum the drawing refers to and measure against it. The witness lines showing what a reading was taken against are drawn on the canvas and included in a PNG export, so the convention is part of the evidence rather than part of the folklore.
This guide is about choosing a measurement direction, not about datum schemes. If your drawing carries a formal GD&T datum reference frame, that frame determines the direction and the answer is not a matter of preference.