Separate source interpretation from mathematical calculation
Every result begins with an interpretation of the source: field mapping, record boundaries, units and reference. The point-file tools preserve original records and use the meanings you explicitly confirm. A coordinate label alone is not sufficient evidence for an axis assignment. For example, Easting and Northing can be exchanged without changing the length of a vector, while the direction and physical position become wrong. Verify asymmetric controls and signed components as well as totals.
A source line identifies where an input record begins; it is not necessarily its point identifier or its sequential index. Quoted multiline text and blank lines explain some differences between these counts. Reports keep source context so a reviewer can return to the correct record. When a calculation cannot use a row, retaining that row and its reason is part of the method, not merely an interface detail or an optional diagnostic convenience.
Record horizontal and vertical units independently
A unit selector normally labels values already present. Explicit conversion is available only where the interface offers that operation. International feet and legacy US survey feet are different units, and a mixed-unit comparison must not be approved simply because the coordinates look similar. Height units need their own confirmation. Horizontal projection conversion preserves height values; it does not transform an ellipsoidal height into an orthometric elevation or apply a geoid model.
Cross-tool handoffs retain source context in the current browser session. A projected result with nonempty height values needs a declared height unit before continuing. KML absolute and relative heights must be confirmed in metres; a ground-clamped export omits altitude coordinates while keeping original values in its report. These checks prevent a change in interpretation from being disguised as a numerical operation. Missing reference information should be resolved from the source rather than inferred from magnitude.
Distinguish exact decimal operations from approximate geometry
Some operations, such as integer identifiers or specific decimal coordinate differences, can preserve exact decimal relationships. Trigonometry, projection algorithms and general geometric calculations still use finite numerical precision. A long displayed number is not evidence of equivalent field accuracy. Conversely, rounding a label for readability should not be confused with the unrounded quantity used internally or retained in a report. Check which output is intended for inspection and which for further computation.
The tools enforce size, range and geometry limits to avoid pretending that unsupported calculations are reliable. A failure at an extreme input should not be bypassed by arbitrarily truncating values. Use the limits shown on the relevant page and retain the original input for investigation. Independent checks should compare results with a tolerance appropriate to the algorithm and numeric scale, while project acceptance tolerances are a separate requirement supplied by the user.
Understand the plane and direction conventions
Ordinary coordinate geometry here uses an Easting-Northing plane. A directed inverse reports changes from the first point to the second. Survey azimuth is measured clockwise from north; a mathematical rotation convention can differ and is selected explicitly where relevant. Horizontal distance is the length in that plane. A slope distance from an instrument needs its selected zenith or vertical-angle reduction before it can be used as the horizontal distance in a forward coordinate construction.
Polygon area concerns an ordered simple planar boundary. A circular-curve schedule follows the declared radius, start, tangent direction and turn. Station/offset calculations use an ordered baseline and require explicit choices where multiple segments are valid candidates. These methods do not automatically become geodesic calculations because a source file originally came from GNSS. Geographic degrees should be processed through an appropriate confirmed projection workflow before applying a metre-based planar construction.
Keep tolerance comparisons and unavailable quantities distinct
In design/as-built comparison, coordinate differences are measured minus design. The horizontal magnitude is formed from the two horizontal components, while the vertical difference retains sign. Separate supplied thresholds evaluate horizontal magnitude and absolute vertical difference. Equality with a threshold is treated inclusively. A missing height produces an unavailable vertical assessment rather than a zero difference or an automatic pass. Unmatched records remain separate from compared pairs.
A file revision comparison asks which records changed; it does not establish construction acceptance. Surface deviation compares a measured height with the design surface at the same plan position, rather than with a point that happens to share a name. Closure, fitting residuals and design deviations also use different reference objects. Do not aggregate their numbers into one universal accuracy label. The relevant project specification determines what evidence and decision rule are appropriate.
Preserve original observations and original surface connectivity
Level and traverse adjustments retain their unadjusted quantities and require an explicit method choice. The equal-weight two-dimensional control fit retains roles, residuals and excluded-control reasons; it does not perform a full covariance-based network solution. Instrument observation reduction uses the angle and height conventions supplied, with no hidden weather, prism or curvature correction. These boundaries make the result reproducible rather than suggesting a more complete survey processing system than is implemented.
For terrain work, the LandXML reader preserves supported original triangle connectivity. Height queries interpolate within those faces. Sections intersect the declared straight baseline or cross lines with original triangles and retain coverage gaps instead of bridging them. Points outside coverage and unresolved geometry remain identifiable. Extracting only vertices is not a lossless surface conversion, and averaging a collection of height differences is not an earthwork volume calculation.
Use examples and regression evidence in the right way
The public examples page supplies small synthetic files with independently understandable outcomes: a rectangle of area six hundred, a three-four-five deviation at a tolerance boundary, and a plane whose height can be checked by substitution. Development tests additionally use documented public point files, instrument observations, official projection definitions and real exported terrain data. Controlled derived cases are labelled as synthetic rather than represented as independent field observations.
Tests examine malformed inputs, empty optional fields, source preservation, unit conflicts, geometry exceptions and exported values, as well as successful calculations. Independent reference outputs reduce the chance that a test merely repeats the implementation's assumptions. Passing these checks supports the stated software behavior; it does not certify an unseen project file or guarantee every browser and operating environment. Recheck a known control after changing a format, reference, input source or output destination.
Keep version and handoff evidence together
Download the complete report when a job needs an audit trail. The source, settings, interpretation, tool version and exceptions explain how the final coordinates were produced. A screenshot of one page of a table cannot replace that record. Local files are not automatically backed up by this site, and opening another tool is a separate interpretation step even when a session handoff avoids retyping the columns.
The recent prelaunch review corrected height-unit propagation and Excel cancellation behavior, and added directory navigation and release-configuration checks. Future changes that affect a method should be described alongside their scope rather than hidden behind a general claim of improved accuracy. Report a suspected discrepancy with a minimal non-sensitive example, the expected quantity and the relevant convention. The goal is a calculation that another person can reproduce, investigate and decide whether to use within the project's actual requirements.
Questions and answers
Does more decimal precision mean a more accurate survey?
No. Numerical precision describes representation and calculation. Survey accuracy also depends on control, observation quality, method and reference information that the browser cannot infer.
Does a small fitting residual certify a datum transformation?
No. Residuals describe the selected model and controls. Independent checks, spatial coverage and the suitability of the reference relationship still need review.
Why does the tool refuse an out-of-surface height?
There is no supported containing triangle at that location. Extrapolating or substituting the nearest vertex would be a different method and could create a misleading design elevation.