Comments on: Brian’s (Purely) Functional Brain http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/ Coding, Mostly Sun, 29 Aug 2010 20:08:13 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Noah Easterly http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-86 Noah Easterly Fri, 04 Jun 2010 21:14:06 +0000 http://blog.willdonnelly.net/?p=293#comment-86 I think your implementation of clip has some bugs: [cce_haskell] > clip max val | val > | val > max = clip max $ val - max + 1 > | otherwise = val > > -- so case by case > clip max (-1) = max - 2 > clip max 0 = max - 1 > clip max 1 = 1 > clip max 2 = 2 > -- ... > clip max (max-1) = max - 1 > clip max max = max > clip max (max + 1) = 2 > clip max (max + 2) = 3 > [/cce_haskell] So when we wraparound to the nonpositive, we never get a value of max, and when we wrap around past max, we never get a value of 1. Fixing it is pretty easy though: [cce_haskell] > clip max val | val > | val > max = clip max $ val - max > | otherwise = val > > -- so case by case > clip max (-1) = max - 1 > clip max 0 = max > clip max 1 = 1 > clip max 2 = 2 > -- ... > clip max (max-1) = max - 1 > clip max max = max > clip max (max + 1) = 1 > clip max (max + 2) = 2 > [/cce_haskell] I think your implementation of clip has some bugs:

>         clip max val | val
>                      | val > max = clip max $ val - max + 1
>                      | otherwise = val
>
>         -- so case by case
>         clip max (-1) = max - 2
>         clip max 0 = max - 1
>         clip max 1 = 1
>         clip max 2 = 2
>         -- ...
>         clip max (max-1) = max - 1
>         clip max max = max
>         clip max (max + 1) = 2
>         clip max (max + 2) = 3
>

So when we wraparound to the nonpositive, we never get a value of max, and when we wrap around past max, we never get a value of 1.

Fixing it is pretty easy though:

>         clip max val | val
>                      | val > max = clip max $ val - max
>                      | otherwise = val
>
>         -- so case by case
>         clip max (-1) = max - 1
>         clip max 0 = max
>         clip max 1 = 1
>         clip max 2 = 2
>         -- ...
>         clip max (max-1) = max - 1
>         clip max max = max
>         clip max (max + 1) = 1
>         clip max (max + 2) = 2
>
]]> By: Noah Easterly http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-85 Noah Easterly Fri, 04 Jun 2010 20:12:00 +0000 http://blog.willdonnelly.net/?p=293#comment-85 I think your implementation of clip has some bugs: <code> clip max val | val max = clip max $ val - max + 1 | otherwise = val -- so case by case clip max (-1) = max - 2 clip max 0 = max - 1 clip max 1 = 1 clip max 2 = 2 -- ... clip max (max-1) = max - 1 clip max max = max clip max (max + 1) = 2 clip max (max + 2) = 3 </code> So when we wraparound to the nonpositive, we never get a value of max, and when we wrap around past max, we never get a value of 1. Fixing it is pretty easy though: <code> clip max val | val max = clip max $ val - max | otherwise = val -- so case by case clip max (-1) = max - 1 clip max 0 = max clip max 1 = 1 clip max 2 = 2 -- ... clip max (max-1) = max - 1 clip max max = max clip max (max + 1) = 1 clip max (max + 2) = 2 </code> I think your implementation of clip has some bugs:

         clip max val | val  max = clip max $ val - max + 1
                      | otherwise = val

         -- so case by case
         clip max (-1) = max - 2
         clip max 0 = max - 1
         clip max 1 = 1
         clip max 2 = 2
         -- ...
         clip max (max-1) = max - 1
         clip max max = max
         clip max (max + 1) = 2
         clip max (max + 2) = 3

So when we wraparound to the nonpositive, we never get a value of max, and when we wrap around past max, we never get a value of 1.

Fixing it is pretty easy though:

         clip max val | val  max = clip max $ val - max
                      | otherwise = val

         -- so case by case
         clip max (-1) = max - 1
         clip max 0 = max
         clip max 1 = 1
         clip max 2 = 2
         -- ...
         clip max (max-1) = max - 1
         clip max max = max
         clip max (max + 1) = 1
         clip max (max + 2) = 2
]]> By: Stephen Mann http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-84 Stephen Mann Tue, 16 Feb 2010 04:51:24 +0000 http://blog.willdonnelly.net/?p=293#comment-84 I hadn't come across xmonad before. It looks awesome! Thanks for sharing. I've been using this (http://stephenmann.net/2009/11/24/vim-is-my-grails-ide/) trick to interact with the terminal from gvim. It's a rip off Jonathan Palardy's SLIME trick (http://technotales.wordpress.com/2007/10/03/like-slime-for-vim/). I've recently been looking into emacs, looking for better editor/terminal interaction than a screen hack. But I've been using vi so long, I don't know if I can (or really want) to make the switch. Maybe xmonad will make switching apps so fast, it won't matter that there's no direct interaction. I hadn’t come across xmonad before. It looks awesome! Thanks for sharing.

I’ve been using this (http://stephenmann.net/2009/11/24/vim-is-my-grails-ide/) trick to interact with the terminal from gvim. It’s a rip off Jonathan Palardy’s SLIME trick (http://technotales.wordpress.com/2007/10/03/like-slime-for-vim/).

I’ve recently been looking into emacs, looking for better editor/terminal interaction than a screen hack. But I’ve been using vi so long, I don’t know if I can (or really want) to make the switch.

Maybe xmonad will make switching apps so fast, it won’t matter that there’s no direct interaction.

]]> By: Will Donnelly http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-83 Will Donnelly Tue, 16 Feb 2010 01:38:02 +0000 http://blog.willdonnelly.net/?p=293#comment-83 I'm not running gvim, I'm running vim with the csapprox plugin to support 256-color themes, and xmonad handles tiling the three xterms for me. I've found that setup tends to be the best for my development workflow. I keep the source code I'm currently working on in the left half of the screen, and then the right side is given over to any other tools I happen to need. I’m not running gvim, I’m running vim with the csapprox plugin to support 256-color themes, and xmonad handles tiling the three xterms for me.

I’ve found that setup tends to be the best for my development workflow. I keep the source code I’m currently working on in the left half of the screen, and then the right side is given over to any other tools I happen to need.

]]> By: Stephen Mann http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-82 Stephen Mann Mon, 15 Feb 2010 23:57:38 +0000 http://blog.willdonnelly.net/?p=293#comment-82 Great job on the awesome program! That's a beautiful screenshot. How did you manage the screenshot -- getting gvim and a terminal running side-by-side like that? I first thought screen with regions, but then you wouldn't get the colors. Then I thought emacs in viper-mode, but then you wouldn't get the ~'s down the left side. I'm looking around for a good haskell IDE, so this question is about more than simple aesthetics. Great job on the awesome program!

That’s a beautiful screenshot. How did you manage the screenshot — getting gvim and a terminal running side-by-side like that? I first thought screen with regions, but then you wouldn’t get the colors. Then I thought emacs in viper-mode, but then you wouldn’t get the ~’s down the left side.

I’m looking around for a good haskell IDE, so this question is about more than simple aesthetics.

]]> By: Brians Brain — Errata & Echoes « BEST IN CLASS http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-81 Brians Brain — Errata & Echoes « BEST IN CLASS Sun, 25 Oct 2009 23:40:05 +0000 http://blog.willdonnelly.net/?p=293#comment-81 [...] Will Donnelly really impressed me. The main logic is only about 6 lines of beautiful Haskell  [...] [...] Will Donnelly really impressed me. The main logic is only about 6 lines of beautiful Haskell  [...]

]]> By: Michael Schade http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-80 Michael Schade Fri, 16 Oct 2009 22:43:24 +0000 http://blog.willdonnelly.net/?p=293#comment-80 Very nicely done. It has me interested in Haskell now (don't let Python know), that's for sure. Very nicely done. It has me interested in Haskell now (don’t let Python know), that’s for sure.

]]> By: silver http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-79 silver Fri, 16 Oct 2009 20:44:29 +0000 http://blog.willdonnelly.net/?p=293#comment-79 Sorry for offtopic, but can you share your haskell-mode? I like your colors :> Sorry for offtopic, but can you share your haskell-mode? I like your colors :>

]]> By: drhodes http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-78 drhodes Fri, 16 Oct 2009 03:25:44 +0000 http://blog.willdonnelly.net/?p=293#comment-78 RE: square world, Oh! I suppose the symmetry has something to do with it. I'm having a great time reading this program. It integrate many different aspects: SDL, Concurrency, Random, Monads and Strategies all at once. Thanks for taking the time to post, it helps. RE: square world, Oh! I suppose the symmetry has something to do with it. I’m having a great time reading this program. It integrate many different aspects: SDL, Concurrency, Random, Monads and Strategies all at once. Thanks for taking the time to post, it helps.

]]> By: Will Donnelly http://blog.willdonnelly.net/2009/10/14/brians-purely-functional-brain/comment-page-1/#comment-77 Will Donnelly Thu, 15 Oct 2009 23:31:10 +0000 http://blog.willdonnelly.net/?p=293#comment-77 drhodes: <br> There's good reason for having independent sizes on the different axes: with a perfectly square world you can end up with two gliders remaining, and the two will never collide, but with some properly selected sizing that can be made impossible. <br> Sebastian: <br> You may be right. I think the list comprehension was actually left over from an earlier iteration wherein I needed a similar comprehension multiple times, and I just never really looked at changing it because it wouldn't have altered the line count, and seems plenty clear to me. drhodes:

There’s good reason for having independent sizes on the different axes: with a perfectly square world you can end up with two gliders remaining, and the two will never collide, but with some properly selected sizing that can be made impossible.

Sebastian:

You may be right. I think the list comprehension was actually left over from an earlier iteration wherein I needed a similar comprehension multiple times, and I just never really looked at
changing it because it wouldn’t have altered the line count, and seems plenty clear to me.

]]>