17 
classes of each influencing factor. 
The production standards themselves are, for all practical purposes, 
stochastic quantities, not only because of the nature of human performance, 
but also in view of the uncertainty in many of the quantitative aspects 
(eg the length of contour lines to be plotted, etc). Their reliability will not 
only increase with greater differentiation of the model employed, but also 
the longer the period used to collect the statistical data. Care should thus be 
taken to ensure that the data collected refer to truly operational conditions, 
so that allowance is also made for unavoidable breaks, difficulties encountered, 
repetitions, etc. It can be assumed that stochastic variations in production 
standards will be fairly small within an organisation, particularly when 
identical processes have to be carried out and when the terrain types are 
fairly similar. In these circumstances, the establishment and application 
of production standards obviously present few difficulties. 
However, in order that optimal decisions can be taken with regard to 
the procedures or equipment to be used in a particular project, it would be 
highly desirable to derive production standards which would also be applicable 
to projects executed in quite different circumstances, such as other types of 
terrain, by other procedures, or even to projects executed by other organi- 
sations. 
Possible applications of production standards in planning operations 
would thus include: 
— PS used in circumstances identical to those in which the data was acquired 
for their derivation (same organisation, same process, same external 
parameters), 
- PS used by the same organisation, but by modified (or other) processes 
and/or with other external parameters, 
— PS used by other organisations but by nearly identical processes and 
external parameters, 
— PS used by other organisations, by other processes and/or with other 
external parameters. 
In regard to the reliability of the production standards, this will 
obviously decrease in the order of the applications suggested above. The 
effect of limited reliability will however be more significant if absolute costs