Comments for shrughes

Comment on Fluff on asynchronous callbacks. by shrughes

shrughes — Thu, 02 Dec 2010 16:31:38 +0000

And it doesn’t avoid mutual recursion.

Comment on Blog Nihilism by shrughes

shrughes — Tue, 30 Nov 2010 07:08:33 +0000

P.S. I think the frequent use of “I” is an appropriate literary technique for this post.

Comment on in-once by shrughes

shrughes — Wed, 07 Apr 2010 17:18:06 +0000

Right.

> My guess is that it’s these uses that make you “miss Haskell”

What makes me miss Haskell (or GHC) is the libraries and datatypes that I’m familiar with, such as MVars, instead of ones I’m unfamiliar with. I also find myself missing Haskell’s type system — yesterday, I caught myself thinking, “Oh, this is an Arrow,” and sometimes I find myself considering, for brief periods of time, going back and rewriting everything in typed-scheme. Not for longer than 0.68 seconds, though. Lazy evaluation and memory leaks are what I miss the least.

Comment on in-once by Eli Barzilay

Eli Barzilay — Wed, 07 Apr 2010 15:02:05 +0000

What Robby meant in the his comment on the other post applies here too. Iteration sequences and generators are both “destroying themselves” because they use state. In some cases it’s possible to have a sequence or a generator that is safe to restart — like walking down a list. If this is the case, then you can simply use a function that returns a generator instead of using the generator directly — and each call to the function returns a fresh sequence or generator. My guess is that it’s these uses that make you “miss Haskell” — and lazy lists (either using promises or using the lazy scheme language that Robby pointed at) do that well. But there are still cases where this is not a good idea — for example, consider the `in-lines’ sequence which reads from a port — there is no way to make that kind of a sequence restartable with the way mzscheme does ports.

(To get *that* to work, ports will need to be implemented in a way that you can grab them functionally — and that leads to potential memory problems where simple mistakes like closing over a newly open port lead to the whole file staying in memory, and it’s also a partial solution since a full one requires doing similar things with random numbers or the other 2000 side-effect pieces in plt.)

Comment on Inelegant Combobulation by shrughes

shrughes — Thu, 01 Apr 2010 18:06:46 +0000

It will work in a mzscheme shell running inside a *shell* buffer inside emacs.

Comment on forzan’s razor by shrughes

shrughes — Thu, 01 Apr 2010 18:05:24 +0000

It’s not a fallacy, it’s a way of life.

Comment on forzan’s razor by forzan

forzan — Thu, 01 Apr 2010 13:34:28 +0000

I require you to fully explain why this should be regarded as a fallacy. Also, you are mixing truth and opinion.

Comment on Inelegant Combobulation by Zombywuf

Zombywuf — Thu, 01 Apr 2010 08:20:41 +0000

Will it work in emacs?

Comment on PLT Scheme sequences are annoying by shrughes

shrughes — Wed, 31 Mar 2010 02:57:03 +0000

Not for what kind of thing?

Comment on PLT Scheme sequences are annoying by Robby Findler

Robby Findler — Tue, 30 Mar 2010 23:16:14 +0000

Sequences are not for that kind of thing. Try putting drscheme into the module language and try running this program:

#lang lazy
(define fibs (cons 1 (cons 1 (map + (cdr fibs) fibs))))