>>33
the problems with programs today are largely problems of misconception. implementation issues like type errors are easy to fix, and simply not that much of a problem in large systems. what really matters is the ability to adapt to changing requirements.
You are still offloading to the machine relatively trivial matters such as checking that the code which was previously expecting a string now works with a tuple, etc. It's not as if there isn't any meaning to types.
static typing is also anti-repl. when you input data into the haskell interpter, it has to decide what it means and it is not always correct.
Yes, I'll have to give you that. The most egregious fault of ghci is its inability to define types at runtime. That's a disgrace. However, you can easily type-hint to solve the problem you cited.
static typing is also anti modular.
That's mostly a buzzword. Yes, it makes coupling apparent and a matter for the compiler, but it isn't as if it weren't there in the first place.
+ for concatenation and addition retardedness
I like my strings/lists and my numbers separate, thank you. It's the one thing that PHP got right. But yes, Haskell does have that kind of polymorphism in Typeclasses. If anything, Haskell already went overboard with genericity with its strings; by being lists it'll choke on decomposed Unicode characters... but at least Everyone's Doing It™.