The leap second should not be commingled with the time we use for appointments.

Frankston, Bob

I come from the world of software and see the need for a stable notation for civil time. For millennia the minute has been considered to represent a portion of the Earth’s rotation. With the advent of atomic clocks and the deployment of GPS we had a need for a more accurate measure of rotation. It seemed necessary to change the definition of the minute to account for the discrepancy.

We need to recognize that we have incommensurate uses for “time”. When we use “time” for scheduling meetings we aren’t concerned with a deep understanding of time – just agreeing on a notation for shared appointments. Implicit ambiguities are common and need to be dealt with as they are recognized. For example, pounds and kilograms have a fixed ratio in day-to-day use but when we leave the surface of the earth we have to distinguish between mass and pounds. In the same way time intervals in databases don’t specify when minutes are recorded and are thus precise but not accurate.

Applications that need to take into account leap seconds can benefit by making their use explicit. The first step is to recognize that the purposes are indeed incommensurate and we would benefit by having a stable way to convert from traditional representations to interval measurements. This provides a solid foundation for variants that are explicit about leap seconds, rotational position and other measures.

For the civil time we need to assure that a minute is always 60 seconds so we can do round trip conversions between TAI and minutes. Today’s databases can’t even represent nonstandard minutes nor can they be supported in standard software libraries.

Return to oral presentation list