Presumably this is newsworthy because everyone assumes building Tetris in Haskell is harder than in an imperative/OO language. This is a good thing because it proves that it's not.
This comes to mind: http://prog21.dadgum.com/23.html
If something non-trivial is required for functional programming, then it is more difficult.
What I can tell you is that the backend (which really needs no state) is much more easier to write in Haskell IMHO.
This says nothing about Haskell or C, but it does say something about me.
I know nothing of your abilities, even so, if haskell is the only language you know well, I'd be impressed if you could make the game as fast as me or most ppl here.
If your first real program in any language was a Tetris game, you'd probably need some help getting started too.
As shown by the various games posted elsewhere in this thread, there are perfectly trivial and straightforward ways to write games in Haskell. If nothing else, Haskell's "do notation" makes it easy to safely mix imperative programming with pure functional programming.
When people do post about writing a game in a new language as a notable achievement, it's usually because they did it in part to learn a new language (as in this case) or a new concept (like using FRP or Erlang-style message passing, instead of imperative programming with mutable state).