chapter  Four
43 Pages

Coding and Information Theory

ByRichard P. Feynman

This chapter discusses the unreliability by considering the aspect of computers for which it is most problematic, that of memory storage. To figure out the actual assignment of coding symbols, people now retrace a path back through the tree. The information is still stored on disks, but individual disks are simply not fast enough to handle the required influx and outflow of data. It is worth examining the practical usefulness of error-detecting codes like this by looking at how the consequences of message failure become less drastic for quite small losses of efficiency. Here is a very simple scheme for checking for single errors, known as a parity checking scheme. An error in the array will then lead to a parity mismatch in both the row and the column in which the error appears, enabling people to pinpoint it precisely. Consideration of a problem will bring people on to consider the famous Sampling Theorem, another baby of Claude Shannon.