6
hitherto been executed by human hand. Analytical aerial triangulation
will exhibits its true merits, we hope, with the increasing utilization
of electronic computer for wider domain of business which in our company
are hitherto treated by numerical procedures. It is our recognition by
practice that almost all the mistakes made by numerical procedure in our
company affect the efficiency and accuracy of our analytical aerial trian<
gulations.
3. General consideration
Process of analytical method consists of measurement and computation.
In our system, the measurement consits of marking of points on diapositives
and observation of the points marked on diapositives. The computation
is, in general, subdivided into some programs according as the individual
organization prefers to be convenient. .Ve have adoped our own system
learned by our own experiences. It is a remarkable fact that this system
which we adopted was quite resembled to t he one which was proposed by
Prof. Van der Weele on the occasion of Milan symposium held on October
19b2. That system was as follows, (Fig.5).
Our system was first organized mainly because our computer has not
sufficient capacity to keep long program at once. But afterward it became
clear that such a system as ours may be most adequate for electronic data
processing of materials similar as of aerial photogrammetry in a factory.
Not in a laboratory but in a factory, it is preferable to depend on not
skillful many persons, and as few skillful operators as possible. And
expecting some mistakes or errors of not skillful operators, it is an
essential point how to settle a system best compatible to compensate the
mistakes or errors automatically and economically. Mistakes or errors
affect the results in different ways according as their characters and
can only be detected in different steps of computations. Vie have learnt
by the practice that in the system illustrated above, the each step of
computation corresponds to the one only in which the mistakes or errors of
individual different character can be detected. Accordingly, we must
expect that in the each step of computation we will find some mistakes or
errors and we will have to make the interruption of computation and after
inspections to make remeasurements. Interruption of computation more or
less involves a waste of computer. Scale of computer must be selected
from the economical standpoint considering the above. Too large content