Information Technology Reference
In-Depth Information
curacy. However, before jumping at the chance to be more accurate,
we need to consider what accuracy we expect or need.
For example, if you are working on a space mission involving a
trip from the earth to Mars, then you may need to know the desired
direction of travel to a considerable degree of accuracy. If the direc
tion were off by even one degree, that one degree over thousands of
miles could result in flying much too far from Mars to obtain orbit,
or too close, which would end in crashing. Similarly, if you are tim
ing runners during a track meet, then an error of 0.1 seconds could
make a difference in who wins a race.
On the other hand, if you are determining how much paint you
need to paint your house, and if house paint comes only in gallon
containers, then computations need be accurate only to the whole
gallon. Further, with variations of paint thickness, limitations on
how accurately you can measure the height of your house, and fluc
tuations in your painting technique, your initial data for computa
tions will be quite rough, and you cannot expect answers to many
decimal places of accuracy if the beginning measurements are not
very accurate.
Accuracy is vital in some contexts, but not in others. Further,
accuracy in answers is limited by accuracy in initial measurements.
We may not want computers to add considerable error to computa
tions, but we should not take time and resources for unneeded ac
curacy, and we should not expect processing power to make up for
errors in initial data.
As with considerations of features, speed, and graphics, we
need our results to be sufficiently accurate to meet our needs; ad
ditional accuracy, however, may provide little benefit. Further, to
obtain greater accuracy, a computer may require additional stor
age space or processing time. Thus, when you read about new
computer systems with new capabilities, your analysis might first
consider the extent to which reported advances will actually help
in solving the problems you care about. You also should consider
associated costs for equipment or processing, and whether new
systems might create new frustrations as new errors are encoun
tered. New or expanded systems can be most helpful in some situ
ations, but of marginal value in others. Your decisions on new sys
tems should depend on careful analysis, not just on appealing
stories or advertising.
Search WWH ::




Custom Search