Dave wrote:
blocks are more basic than lambdas
The biggest difference is that lambda syntax (being functional) almost always expect an argument; a function with no arguments doesn't make any sense. Block syntax is optimised to work well in the no-argument case.
in Scheme cond is a primitive, even though it could have been done differently
the questions is whether you quote at definition side (Scheme) or use side (Smalltalk)
Chris Double (hey two Kiwis in the same thread :-) wrote:
This is how the Pony programming language treats closures. They are sugar for objects with an apply method:
yep, lots of langauges do that, incl. Java. The catch is: where does "return" return to?
J
IMO, the biggest difference is the syntactic weight: "lambda" vs "[]".
Yes, the lack of a non-local return prevents the use of closures for control structures in many languages. The "maybe" monoid is an interesting alternative that could be viewed as a partial workaround in functional languages. Exceptions, too, sorta.
I've not looked at Pony yet.
- David (from iPad, typos likely)
On Jun 27, 2016, at 6:44 PM, James Noble kjx@mcs.vuw.ac.nz [self-interest] self-interest@yahoogroups.com wrote:
Dave wrote:
blocks are more basic than lambdas
The biggest difference is that lambda syntax (being functional) almost always expect an argument; a function with no arguments doesn't make any sense. Block syntax is optimised to work well in the no-argument case.
in Scheme cond is a primitive, even though it could have been done differently
the questions is whether you quote at definition side (Scheme) or use side (Smalltalk)
Chris Double (hey two Kiwis in the same thread :-) wrote:
This is how the Pony programming language treats closures. They are sugar for objects with an apply method:
yep, lots of langauges do that, incl. Java. The catch is: where does "return" return to?
J
self-interest@lists.selflanguage.org
-
David Ungar -
James Noble