The proceedings of the meeting of IAU Commission 31 point out that
The distinction between time epoch and time interval is not clear to everyone.In effort to remedy that situation W. Markowitz, H.M. Smith, and L. Essen introduced an explanatory note on time.
These three noteworthy individuals have stellar biographies
related to the subject of time.
L. Essen was the co-developer of the first cesium atomic
chronometer at the UK NPL.
W. Markowitz was the Director of Time at the USNO, had worked
on the creation of Ephemeris Time, had been chair of IAU Comm.
31 when Essen announced the cesium chronometer, and immediately
acted to perform the calibration of astronomical time scales
with the cesium atomic time scale.
H.M. Smith worked at the Greenwich Observatory Time Service and
spent the 1970s working with the CCIR to get other agencies to
adopt the new UTC with its leap seconds.
This explanatory note was incorporated into the resolutions of the General Assembly (see page 16). Here is an English translation. Below is the French.
These notes revisited, clarified, and emphasized the implications of the original decision to change the radio broadcast time signals which had been made by the IAU in 1955. Those 1955 decisions are available here.
These notes employ terms of art which were commonly understood by astronomers in 1964. The meanings of those terms deserve elaboration for the sake of folks in the 21st century.
In astronomy the term epoch means any calendar date and time which is used as the point of origin for a measurement or series of measurements. Any date and time could serve as the epoch for calculating the position of a celestial body from a series of observations. For a body in the solar system the epoch was a date and time (often a Julian date which was evenly divisible by 5 or 10 days) at which the orbital elements were calculated and from which the motion would be computed to find that body at another date. For a star catalog the epoch was a date and time (usually a Besselian year like B1900.0 and B1950.0 or a Julian year like J2000.0) specifying the celestial coordinate reference frame in which the positions of the printed text had been expressed, and to which the corrections for precession, annual aberration, and proper motion of a star would be added to convert from its catalog position to its position of date.
An important aspect of an epoch was that everyone could agree on when that epoch happened. As of 1964 the only way of specifying a date and time was by using the rotation of the earth to construct a calendar and set a clock. The ability to agree on such an epoch was accessible to anyone who could look up and see the sky, and that remains true so long as anybody keeps track of counting days. In 1964 it was still unclear that it was possible to construct a time scale based on atomic chronometers.
It remains the case that an atomic time scale can only be measured by relatively few highly specialized laboratories performing measurements of cesium and other atoms which have no connection to the rotation of the earth. An atomic time scale cannot provide an epoch of its own. The epoch of an atomic time scale can only be established by linking to some generally visible event, and keeping track of an atomic time scale requires continuous operation of complex equipment and telecommunications.
A dictionary definition is
Point précis, moment exact ou défini, dans le temps.which is to say, the same as epoch above.
The form of Universal Time constructed by measuring the rotation of the earth and applying corrections for seasonal variations in that rotation. In 1964 UT2 was still the recommended basis for legal time, and that had been the case since it was defined in 1955 and implemented starting in 1956. (See the documents from 1955 linked just above.)
This explanatory note is a followup to the IAU decision in 1955 when they voted to redefine the duration of one second to be unrelated to the duration of one day.
During the five years subsequent to 1964 the characteristics of the radio broadcast time scale were re-discussed at many more meetings. Despite warnings from astronomers, in 1970 the CCIR decided to implement a new compromise. With no foresight for the problems that would ensue, the leap second was instituted in 1972.
During the decade after the CCIR implemented the leap second in the radio broadcasts they renamed the time scale UTC and promoted the use of the time scale for all applications. As a result UTC with leap seconds was adopted as the perfectly commendable solution by the CCDS, CGPM, IAU, ITU, CCITT, and national governments.
Unfortunately the POSIX committee did not understand the implications of the leap second and considered that precise time was not important for computing. They disregarded advice to allow the leap second to be counted as part of the time scale used in computers, yet they still chose to specify that computers should use UTC. These failures to communicate and comprehend now cause trouble at each successive leap second.
For more discussion of how different agencies made different choices about the priorities of keeping time see Three desirable characteristics for systems handling time.