I know that was sarcasm (as was my original post), but that's kind of my point. Haskell is a simpler language than Ruby in a lot of ways. But the extra complexity in Ruby (or Perl, whatever -- pick your workaday language) is driven by clear practical concerns.
Where a functional nut would sit and worry about getting a design right that correctly abstracts the properties of any right triangle, a web developer is happy to hard code the values for 3:4:5 because that's all that's required. And he isn't wrong to do so.
Perhaps, but he isn't necessarily right to do so either. If the requirements for the triangle are uncertain enough or change often enough taking the time to get the abstraction right can be appropriate engineering tradeoff.
Where a functional nut would sit and worry about getting a design right that correctly abstracts the properties of any right triangle, a web developer is happy to hard code the values for 3:4:5 because that's all that's required. And he isn't wrong to do so.