Jacob Carlborg:
Why not use proper lambdas instead of strings?
Mostly for personal reasons: quoted strings are sometimes a little shorter, and they require you to use arguments with default names (as "a" and "b"), this increased standardization makes me read them a little faster than lambdas that have varying argument names. If I see q{a + b} I don't need to go look what a and b are, for me it's a "standard" sum, like one to be used by a reduce().
It's a bit like the "pointfree" style in Haskell (http://www.haskell.org/haskellwiki/Pointfree ), also named "tacit" in other languages, where you don't name input variables; that if overused makes the code unreadable, but if used a bit and wisely allows you to underline the transformations done on the data instead of on the (sometimes arbitrary) names you give to parts of generic data.
Bye, bearophile
