For and overview and related topics, see Introduction to PEXlib.
PEXlib uses the floating-point format native to the architecture of the client machine (i.e., the machine on which the application executes). PEXlib expects all of the floating-point values to be in that native format, with a few exceptions for functions which allow the application to specify the floating-point format. These functions, which enable the application to specify a floating-point format, deal primarily with application data that is already in protocol format, or data that the application wants to be encoded into protocol format for its own use.
If the server to which PEXlib is sending data does not support the floating-point format native to PEXlib, then PEXlib chooses an appropriate format which is supported by the server and converts data into that format when formatting protocol to send to that server. In cases where PEXlib must choose a non-native floating- point format, the application can determine the format chosen by invoking the PEXGetProtocolFloatFormat function. This is useful with functions dealing with protocol formatted data. It helps avoid unnecessary format conversions when sending the data to the PEX server.