Question:
I was told that in order to make my program easier to read, I should
use
try to cover all of my program, then catch every exception at the end
of the program. Does this make any sense?
Answer:
You should catch exceptions where you intend to handle them. It makes
no sense to catch exceptions only at the end of a program when many
exceptions are recoverable. For example, network servers need to be
able to recover from network I/O errors. You can only do this if you
catch exceptions at the point of an anticipated failure point. But
you must ensure that you take proper fault-recovery steps. Only fatal and
non-recoverable exceptions resulting in the termination of the program
should be caught at the end of the program.