Methodology

Distance

Distances are geodesic — the shortest path over the surface of the Earth — computed with Vincenty's inverse formula on the WGS-84 ellipsoid, the same reference ellipsoid GPS uses. Its defining parameters are a semi-major axis of 6,378,137 m and a flattening of 1/298.257223563.

The simpler alternative is a spherical great-circle calculation on a mean radius of 6371.0088 km. This site does not use it for displayed distances, because the Earth is flattened at the poles and a spherical model runs short on high-latitude routes — about 15 km short on JFK–LHR, or 0.27%. That is enough to disagree with every published figure a visitor might check the result against. The ellipsoidal calculation agrees with them.

The spherical formula is still used internally where a fraction of a percent cannot change the answer, such as ranking nearby airports by proximity. Vincenty's formula does not converge for near-antipodal points; in that case the spherical result is used instead.

Miles are km ÷ 1.609344; nautical miles are km ÷ 1.852. Both are exact definitional conversions.

What this does not account for:

• Aircraft do not fly geodesics. They follow published airways, avoid restricted airspace, route around weather and step around terrain. Real tracks typically run a few percent longer, and considerably longer where airspace is constrained.
• Coordinates are the airport reference point, not the runway threshold or the gate. On a large airport this is a difference of a kilometre or two.
• Altitude is ignored. The distance is measured at the ellipsoid surface, not at cruise level.

Flying-time estimates

Block time is estimated from distance alone:

hours = distance_km / 840 + 0.58

That is a cruise speed of 840 km/h plus a fixed 35-minute allowance for taxi, climb, descent and approach. It is deliberately simple, and it is deliberately labelled an estimate everywhere it appears.

What this does not account for:

• Winds. The jet stream makes eastbound transatlantic and transpacific flights materially faster than the same route westbound. This is the largest source of error, and it is directional — the same estimate is too long one way and too short the other.
• Aircraft type. Cruise speeds vary meaningfully across types.
• Airport congestion. Taxi times at a large hub can exceed 35 minutes on their own.
• Routing. Longer actual tracks mean longer actual times.

These figures should never be used to judge whether a connection is feasible. Use the airline's published schedule.

Time zones and arrival dates

Each airport is assigned an IANA time zone — Europe/London, America/New_York — resolved from its coordinates at build time. Offsets are then computed with the browser's own Intl implementation for the specific date in question, not for today.

That last detail matters. Because the offset is resolved per date:

• Daylight saving is handled correctly, including the weeks when two regions have switched on different dates and the usual gap between them is temporarily wrong
• Zones that are not whole hours from UTC — India at +5:30, Nepal at +5:45, Chatham Islands at +12:45 — work correctly
• Historical and future rule changes are reflected as far as the visitor's browser data goes

Arrival time is computed by:

1. Interpreting the departure date and time as wall-clock time in the origin's zone and converting to a UTC instant
2. Adding the estimated flying time
3. Rendering that instant as wall-clock time in the destination's zone
4. Comparing the two calendar dates to determine the day shift

Because the arrival time inherits the flying-time estimate, it carries the same uncertainty. The date shift it reports is nevertheless reliable, since an hour of error rarely changes which calendar day you land on.

Date-line detection

A route is flagged as crossing the International Date Line when the raw longitude difference between the two airports exceeds 180°, meaning the shorter great-circle path passes through the ±180° meridian. This is a geometric test on the airport coordinates, and it is independent of the time-zone calculation above — which is why a flagged route and a reported day shift are two separate pieces of evidence about the same journey.

The airport dataset filter

The source dataset contains roughly 86,000 records including heliports, seaplane bases, disused airfields and private strips. This site narrows that to 4,133 airports by keeping records that:

• Have a valid three-character IATA code
• Are marked as having scheduled commercial service

The 50 airports with planned entity pages are retained regardless. Where two records share an IATA code, the entity record wins, then the larger airport type.

The trade-off is deliberate: a smaller index means faster search and a smaller download, at the cost of excluding airports that have no scheduled service. If you are looking for a general aviation field, this is not the right database.

Content review

Guides covering regulatory topics — proof of onward travel, visa document requirements — carry a "requirements last reviewed" date and link to official sources. These are reviewed quarterly and whenever a significant change comes to our attention. Guides explaining stable concepts, such as how a PNR works, are updated when the underlying practice changes.

Requirements are described conditionally throughout, because they genuinely vary by nationality, route and category. Where a guide says "check with your consulate", that is not hedging — it is the accurate answer.

Corrections

If a calculation, a data record or an explanation here is wrong, tell us and it will be corrected. See also data sources.