I can think of another example of where relocatable code can be valuable in a ROM/EPROM based system. There are many 8051 based system designs where the program size far exceeds the 64K address limitations of the original Intel architecture. Many of these systems use a scheme known collectively as "bank switching" to permit multiple blocks or pages of code to be utilized. It is easy to anticipate that some bank switching schemes may allow multiple memory windows to the paged memory at two or more different addresses in the 8051 memory map. If code written for inclusion in a bank switch page is written in the "relocatable format" then that code will be able to run when its page is mapped to either of its multiple bank acess windows.

Michael Karas