Part of the computational model is the environment in which your application executes. The environment contains state information you control. The rounding mode, as defined by the IEEE floating-point standards, determines how mathematical results not exactly representable are coerced to the destination format. Rounding may be to nearest (the default), toward zero, toward 
, or toward 
. The floating-point standards also specify a set of five flags associated with the possible exceptions: inexact (result), overflow, underflow, divide-by-zero, invalid (operation).
Once set, these flags remain set until you clear them. The CommonPoint system provides a set of classes for manipulating the environment so you can handle rounding modes and exceptions and so you can write robust functions.
[Contents]
[Previous]
[Next]
Click the icon to mail questions or corrections about this material to Taligent personnel.