DMS and decimal degrees converter

An angle in surveying is written in degrees, minutes and seconds, and it has to be in decimal degrees before any trigonometry can touch it. That conversion is trivial arithmetic and a reliable source of error, because the notation is mixed-radix, because everybody types it differently, and because the rounding at the end has a trap in it that most spreadsheet formulas fall straight into.

The sexagesimal system, and why it survives

A degree is divided into sixty minutes of arc, and a minute into sixty seconds. The system is Babylonian, it is older than the decimal point, and it has outlasted every attempt to replace it because sixty divides evenly by two, three, four, five, six, ten, twelve, fifteen, twenty and thirty. A third of a degree is exactly twenty minutes. A third of a decimal degree is 0.333… forever.

That last point catches people. A negative angle of −34°12′45″ is minus the whole thing: the minutes and seconds are magnitudes, not signed contributions. Writing it as −34 + 12/60 + 45/3600 gives −33.7875°, which is wrong by nearly half a degree, and the sign of the error flips depending on which side of zero the angle sits.

The notations surveyors actually type

There is no standard keyboard notation for an angle, so a working converter has to accept all of the ones in circulation. Every form below is read identically by the fields above.

Two rules keep that flexibility from becoming laxity. Minutes and seconds must lie below sixty, so an entry of 90 61 00 is rejected rather than quietly carried to 91°01′00″ — a keying error is cheapest to catch at the keyboard. And a field with more than three numeric groups is rejected outright, because it is not an angle at all.

The 59′60″ rounding bug

Converting decimal degrees back to DMS looks like three lines of code: take the whole degrees, multiply the remainder by sixty for the minutes, multiply that remainder by sixty for the seconds, and round the seconds for display. Written in that order it is wrong, and it is wrong exactly at the values you are most likely to be checking.

Take 12.9999999°. The whole degrees are 12. The minutes work out to 59, and the seconds to 59.99964. Round the seconds to two decimals for display and you print 12°59′60.00″ — an angle with sixty seconds in it, which does not exist. The naive code produced a display value that violates the very system it is displaying.

The same trap moves as you change the seconds precision, which is why the decimals selector matters more than it looks. An angle that displays cleanly at two decimals can carry at zero decimals and not at three, so the boundary case is not one fixed number — it is a whole family of them, one per precision setting.

Radians and gradians

Radians are the natural unit of angle: the arc length subtended on a unit circle. Every trigonometric function in every programming language takes radians, which is why the conversion sits between your DMS value and any computation you do with it.

Gradians appear in European instrument settings and in some continental survey software, where a right angle of exactly 100 gon makes quadrant arithmetic clean. An instrument accidentally left in gon reports a 90° angle as 100, which is close enough to a plausible number that it can survive a whole day of observations before anyone notices.

The arc second and the constant ρ″

Error analysis in surveying is done in seconds of arc, because that is the resolution instruments are specified at and the unit angular closures are reported in. The bridge between an angular error and the linear error it causes is one constant.

That approximation is the small-angle relation, and within the range of angles that survey errors occupy it is exact for every practical purpose. It is what lets you answer the question that actually matters on site: a 5″ direction error over a 900 ft sight line displaces the point by 900 · 5 / 206,265, which is 0.022 ft. Two hundredths of a foot, and therefore not the thing that ruined your closure.

Running the same arithmetic backwards is how a specification gets set. If a layout point has to land within 0.03 ft at 600 ft, the direction has to be good to 0.03 · 206,265 / 600, which is about 10″ — and that number tells you what class of instrument the job needs before anybody drives to it.

Significant figures in an angle

Precision in an angle is not counted in digits the way it is in a distance, because the units are not decimal. One second of arc is about 4.85 microradians, or roughly five parts per million. A distance quoted to five parts per million is 1,000.005 ft — seven significant figures. So an angle read to the nearest second carries about as much information as a distance read to the nearest hundredth of a foot over a thousand feet, and the two observations are well matched, which is why a total station specified at 5″ is usually paired with distance work at a hundredth.

The practical consequence is a display rule. Printing an angle to three decimals of a second implies a precision of about five parts per billion, which no field instrument delivers and no closed traverse justifies. Carry the extra digits through the computation, and print the number to the precision the observation actually had.