get rid of unary '-' operator

[ijones]

See NegativeSyntax.

[guest]

Pros:

• Removes an embarrassing special case from the grammar.
• Makes the section (- x) work as expected.
• Expressions such as -123 would not require parentheses (foo (-1) could be written simply as foo -1)
• Expressions such as -1 would be permitted in k-patterns even if n+k patterns end up being thrown out.
• You can say negate $ 1 + 2 if you don't like parentheses.
• Some may argue that it is a bad idea to give lexical significance to the space between - and its argument lexically, that is, to give (- 1) and (-1) a different meaning. However, the precedent for reusing an infix operator symbol as a part of a symbol has already been made in two places in the Haskell 98 standard, namely in the cases of . operator (which is also used to separate module components) and +/- in the floating point notation (1E-5). Further, arguably expressions such as - 123 (with a space) seldom appear in real-life code.
• Other extensions (unboxed types) already append an operator symbol to identifiers: giving (- 1) and (-1) different meanings is no more confusing than having different meanings for (Int#) and (Int #)

Cons:

• Expressions of the form -x (where x is not a constant) need to be rewritten as negate x, although the later is arguably much more cleaner, especially since, more often than not, x is a complex paranthesized expression rather than a simple variable name.
• Possible confusion to the beginners (we can write -1 but not -x.) However, I think that the strange behaviour of sections and the need for parentheses around -1 is already confusing enough to beginners, and therefore this change would actually make Haskell easier to learn, not harder. negate x looks so much more like ordinary Haskell code!