All engineers, regardless of the branch of the profession to which they belong, are constantly faced with the problem of measurement. It may be of time, mass, force, temperature, the flow of an electric current, length, angle, and so on; or it may be of the effects of some of these in combination. Almost invariably, the results of such measurements will determine the course of action the engineer takes thereafter. Thus the results obtained by measurements provide information upon which decisions are made.

All such measurements form part of the science of metrology. The mechanical and production engineer are, however, especially concerned with the measurement of length and angle. Of these, length is of fundamental importance since angular measurement may be carried out by the appropriate use of linear measurements in combination.

Thus the purpose of any measurement is to provide a service to enable a decision to be made. The service will not be complete unless the measurement is made to an acceptable degree of accuracy, but it must be realized that no measurement is exact. It is therefore necessary to state not only the measured dimension, but also the accuracy of determination to which the measurement has been made. As far as possible the errors inherent in the method of measurement used should be kept to a minimum, and having minimized the error, its probable magnitude, or accuracy of determination, should be stated.

It follows that it is not enough to state that the nominal size of a gauge block is, for example, 30mm. It is also necessary to state:

(a) The measured error in the block,

e.g. - 0,000 2mm

(b) The accuracy of determination,

e.g. +/- 0,000 4mm

The user, having this information, may now avail himself of it if necessary. If the gauge block is used to set the datum for a vernier height gauge which can only be read to 0,02mm, then the gauge block errors are negligible and can be ignored. If on the other hand, it is used to set up a comparator whose scale divisions represent 0,001mm, then the measured error is significant and must be considered, and the accuracy of determination of the gauge block must be incorporated in the accuracy of determination of the comparison being made.