Programming C# 5.0

Chapter 8. Exceptions

Some operations can fail. If your program is reading data from a file stored on an external drive, someone might disconnect the drive. Your application might try to construct an array only to discover that the system does not have enough free memory. Intermittent wireless network connectivity can cause network requests to fail. One widely used way for a program to discover these sorts of failures is for each API to return a value indicating whether the operation succeeded. This requires developers to be vigilant if all errors are to be detected, because programs must check the return value of every operation. This is certainly a viable strategy, but it can obscure the code; the logical sequence of work to be performed when nothing goes wrong can get buried by all of the error checking, making the code harder to maintain. C# supports another popular error handling mechanism that can avoid this problem: exceptions.

When an API reports failure with an exception, this disrupts the normal flow of execution, leaping straight to the nearest suitable error handling code. This enables a degree of separation between error handling logic and the code that tries to perform the task at hand. This can make code easier to read and maintain, although it does have the downside of making it harder to see all the possible ways in which the code may execute.

Exceptions can also report problems with operations where a return code might not be practical. For example, the runtime can ...