On 3/22/14, 12:43, Russel Winder wrote:
On Sat, 2014-03-22 at 16:14 +0000, Brian Rogoff wrote:
On Saturday, 22 March 2014 at 13:03:06 UTC, Russel Winder wrote:
ALGOL60 did not have significant whitespace and an offside
rule, just
like C, C++ and D don't, whereas Python, OCaml, etc. do.
I've programmed in OCaml for many years and I somehow missed the
significant whitespace. Even the Revised syntax for OCaml (the
improved and unused one) did not use significant whitespace,
though I recall that there were unloved projects to provide such
a syntax.
I appear to have typed OCaml when I meant Haskell, possibly because I am
trying to build Unison. You are correct (obviously :-) OCaml does not
use an offside rule approach. In his response, Paulo points out that F#
does, I did not appreciate this, so that is definitely a WILT.
C++ has a much nastier syntax than D (IMO of course :-) but the
SPECS proposal for a resyntaxed C++ never caught on. I liked some
of the improvements suggested there, in particular the more
Pascal-ish or Scala-ish declaration syntax, and would have liked
something like that in D, but there are so many more issues to be
fixed that daydreams of improved syntax seem frivolous to me.
The Scala, Go, Rust, etc. use of "type after variable name" reads better
for me, but C, C++, D, Java, Groovy, Ceylon are all "type before
variable" (well the C++ rule is spiral out, but…), so I just get used to
switching.
I find I think of the type as an adjective, and since I'm only fluent in
english it makes perfect sense that the "adjective" would come before
the "noun".
What is X? X is an integer. Integer describes what X is.
"type after variable name" just doesn't have that mental model to it,
hence I like it less.
