[Haskell-cafe] Haskell Propeganda

[Tim Newsham]

> As a day-time java programmer, I can say from experience that sometimes (100% 
> pure) Java programs DO segfault.
>
> I've had it happen to me, and while you can justifiably say it's an error in 
> the JVM somehow triggered by your program behaviour/timing, that doesn't help 
> you very much at the time.

Sure, and just as java has its runtime, so does Haskell.  Runtimes
are software and prone to being buggy.  Runtimes are often written
in languages like C where bugs can lead to memory corruption, crashes
and arbitrary malicious code execution.

Here's a small example that I unfortunately ran across today:

ghc -c wxdirect/src/CompileClassTypes.hs -o
   dist/wxdirect/CompileClassTypes.o -idist/wxdirect -odir dist/wxdirect
   -hidir dist/wxdirect  -package parsec -cpp  -package time
Segmentation fault (core dumped)

:(

> Maarten Hazewinkel

The point being that Haskell, while having a theorem checker that helps
programmers find and avoid bugs and while being based on semi-formal
concepts that can be used to avoid some pitfalls, is still no silver
bullet against any and all crashes.  Promising would-be converts that
it is will only lead to disappointment.

Tim Newsham
http://www.thenewsh.com/~newsham/