First of all, I never claimed that libraries didn't use records; certainly nearly all do. I was talking about lenses when I said you don't need to master them t

First of all, I never claimed that libraries didn't use records; certainly nearly all do. I was talking about lenses when I said you don't need to master them to write good code. I'm not sure what it means for a record system to be "pathetic", but although there are issues with records in Haskell, it (a) is improving with new features forthcoming in GHC and libraries, and (b) has not stopped developers from writing millions of lines of robust, practical code. I think the "record problem" in Haskell is a legitimate one but not one which will ever prevent you from writing good code. At worst it's an annoyance. As to the questions about fixity, this is once again in practice not a big problem. First of all the notions of operator fixity and association are familiar to all of us since grade school. We all know "please excuse my dear aunt sally" or some similar mnemonic for the precedence rules of the operators of arithmetic. It's this rule that lets us write "a^2 + 2 * b^2 * c^(2 / 3)" instead of "(a^2) + ((2 * (b ^ 2)) * (c ^ (2 / 3)))". In other words it's a way to make an expression easier to read and parse by removing superfluous parentheses. It's not arbitrary at all, and in fact operator precedence rules is a feature that every other mainstream language has. As to what semantic benefits there are to fixity, fixity is a purely syntactic notion. It has nothing to do with semantics. Giving your operator an associativity and precedence only affects how a particular expression is parsed, not how it's interpreted or evaluated. I'd like to mention that your comments so far seem unnecessarily combative and hyperbolic. You use phrases and words like "pathetic", "good luck", "useless flexibility", "zero benefits", "please enlighten me". This injects contention into a discussion where there needn't be any. Furthermore, your statement that you only use Haskell for pure computations with minimal IO makes me question the legitimacy of your criticisms. Certainly, there are countless others who have written all manner of heavily interactive programs, web servers, database interaction, or IO-intensive programs in Haskell. If Haskell isn't your cup of tea, that's fine, but do you really think you're in a position to be leveling such criticisms when your own usage is so restricted?