In FP, I think this sort of problem is generally handled via algebraic data types rather than exceptions. In particular this directly addresses the issue of "exceptions don't necessarily shout themselves out", since the compiler warns you if you've missed a case.

They sound mathy, but algebraic data types are actually a pretty simple concept. I think the "Learn You a Haskell" explanation is decent: http://learnyouahaskell.com/making-our-own-types-and-typeclasses

Provided I understand the context properly, actually using exceptions for this sort of issue would be extremely rare practice.

On Fri, Mar 29, 2013 at 12:21 AM, luc taesch <luc.taesch@gmail.com> wrote:

I was looking for some link introducing the way FP/ Haskell handles errors and Exceptions.

This is for a non FP Guy, and ideally withought scaring them with Monads and category theory :-).

for the background :

the guy said : As I mentioned in another thread in banking (in particular) it is the exception processing that often dominates the functionality of a system - as the core concept is generally very straightforward. Developing for "exception handling" (not in a Java/C++ sense) is a tricky thing - as the exception don't necessarily shout themselves out - and are often why we have large misunderstood legacy systems which are hard to replace.

_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe