On 23 Aug 2008, at 22:36, Matus Tejiscak wrote:
On So, 2008-08-23 at 22:16 +0200, Thomas Davie wrote:
Today I made an interesting discovery.
We all know the benefits of a strong type system, and often tout it as a major advantage of using Haskell. The discovery I made, was that C programmer don't realise the implications of that, as this comment highlights:
http://games.slashdot.org/comments.pl?sid=654821&cid=24716845
Apparently, no one realises that a SEGFAULT is a type error, just not a very helpful one.
Bob
Type errors are useful because they emerge at compile time and prevent you from compiling (and running) a broken program. A segfault is a runtime error and as such provides no such guide -- it may or may not arise and you don't know something's wrong until sigsegv kills your app, screws all your data, crashes the airplane etc. (without the possibility to tell whether/when it will happen).
I guess I didn't express my point very clearly... That C programmers apparently don't realise that a type system that's sound will give them something -- i.e. their programmer won't ever segfault. I wonder when we try to advertise Haskell if we should be saying "we can give you programs that never segfault", instead of "we have a strong type system". Bob