News for hackers. Stuff that matters.
Re:Can Programming Be Liberated from the von Neumann Style? ... by eru on Tuesday September 08, 2015 @03:07AM (#10184389)
For most tasks we can restrict ourselves to a terminating subset without bottom. See Agda, Coq, Isabelle, etc.
Re:Can Programming Be Liberated from the von Neumann Style? ... by eli_gottlieb on Tuesday September 08, 2015 @08:09AM (#10185186)
Bingo. But most people also like to kvetch and moan about the restrictions imposed by programming with only structural recursion and productive corecursion -- even though those are usually exactly what we want.
Re:Can Programming Be Liberated from the von Neumann Style? ... by eru on Tuesday September 08, 2015 @11:23PM (#10189496)
Especially if you get an easy way out for when you need it.

Like unsafePerformIO, but for calling not-proven-to-be-terminating functions. Or an explicit marker like the wrapping monads for side-effecting functions.

Re:Can Programming Be Liberated from the von Neumann Style? ... by dragonwriter on Tuesday September 08, 2015 @05:26PM (#10188301)
> For most tasks we can restrict ourselves to a terminating subset without bottom. See Agda, Coq, Isabelle, etc.

While those are interesting languages, I kind of was under the impression that they were generally not seen as suitable for "most tasks".

Re:Can Programming Be Liberated from the von Neumann Style? ... by eru on Tuesday September 08, 2015 @11:21PM (#10189492)
I mean, even in Haskell, big chunks of every program can be automatically recognized as `safe Haskell'.