scheme递归
主要参考:
http://www.shido.info/lisp/scheme7_e.html
Function fact that calculates factorials.
(define (fact n)
(if (= n 1)
1
(* n (fact (- n 1)))))
(fact 5) is calculated like as follows:
(fact 5)
⇒ 5 * (fact 4)
⇒ 5 * 4 * (fact 3)
⇒ 5 * 4 * 3 * (fact 2)
⇒ 5 * 4 * 3 * 2 * (fact 1)
⇒ 5 * 4 * 3 * 2 * 1
⇒ 5 * 4 * 3 * 2
⇒ 5 * 4 * 6
⇒ 5 * 24
⇒ 120
(fact 5) calls (fact 4), (fact 4) calls (fact 3), then finally (fact 1) is called. (fact 5), (fact 4) ,.., and (fact 1) are allocated at different memory spaces and(fact i) stays there until (fact (- i 1)) returns a value, which wastes the memory space and takes more calculation time because of the overhead of function call.
However, recursive functions can express repetition in a simple manner. Further as lists are defined recursively, lists and recursive functions fit together. For instance, a function that makes all list items twice is defined like as follows. The function should return an empty list if the argument is an empty list to terminate the calculation.
(define (list*2 ls)
(if (null? ls)
'()
(cons (* 2 (car ls))
(list*2 (cdr ls)))))
3. Tail Recursive
Ordinary recursive function is not efficient because of wasting memory and function call overhead. On the contrary, tail recursive functions include the result as argument and returns it directory when the calculation finishes. Especially, as Scheme specification requires conversion of a tail recursive to a loop, there is no function call overhead.
[code 2] shows a tail recursive version of function fact shown in [code 1].
[code 2] fact-tail, tail recursive version of fact
(define (fact-tail n)
(fact-rec n n)) (define (fact-rec n p)
(if (= n 1)
p
(let ((m (- n 1)))
(fact-rec m (* p m)))))
fact-tail calculates factorial like as follows:
(fact-tail 5)
⇒ (fact-rec 5 5)
⇒ (fact-rec 4 20)
⇒ (fact-rec 3 60)
⇒ (fact-rec 2 120)
⇒ (fact-rec 1 120)
⇒ 120
As fact-rec does not wait the result of other functions, it disappears from the memory space when it finishes. The calculation proceeds by changing argument of fact-rec, which is basically the same as loop. As mentioned previously, as Scheme convert a tail recursive to a loop, Scheme can do repetition without syntax for looping.
4. Named let
The named let is available to express loop. [code 3] shows a function fact-let that calculates factorials using named let. The fact-let uses a named let expression (loop), instead of fact-rec shown in [code 2]. First it initializes parameters (n1, p) with n at the line marked with ; 1. These parameters are updated at the line marked with ; 2 after each cycle: Subtracting n1 by one and multiplying p by (n1-1)
A named let is a conventional way to express loops in Scheme.
[code 3]
(define (fact-let n)
(let loop((n1 n) (p n)) ; 1
(if (= n1 1)
p
(let ((m (- n1 1)))
(loop m (* p m)))))) ; 2
5. letrec
While it is similar to the named let, a name defined by letrec can refer itself in its definition. The letrec syntax is used to define complicated recursive functions. [code 4] shows a letrec version of fact.
[code 4]
(define (fact-letrec n)
(letrec ((iter (lambda (n1 p)
(if (= n1 1)
p
(let ((m (- n1 1)))
(iter m (* p m))))))) ; *
(iter n n)))
As shown at the line of ; *, the local variable iter can refer itself in the definition of iter. Syntax letrec is a conventional way to define local functions.
scheme递归的更多相关文章
- Teach Yourself Scheme in Fixnum Days 6 recursion递归
A procedure body can contain calls to other procedures, not least itself: (define factorial (lambda ...
- Scheme中一些函数在C++里面的实现与吐槽
最终我失败了,这是显而意见,我试图在一个很看重类型是什么的语言中实现无类型操作,事实上,哪怕我实现了基本的cons,car,cdr,list后面的代码也无法写下去.比如说list-n,根据 ...
- 与Scheme共舞
发表在<程序猿>2007年7月刊上.不log上写帖子不用考虑版面限制,所以这里的帖子比发表的啰嗦点.赵健平编辑,Jacky,和刘未鹏都给了我非常多帮助,在这里一并谢了.免费的Scheme实 ...
- 递归转手工栈处理的一般式[C语言]
是任意形式的递归,是化解的一般式. 主题所谓的“递归调用化解为栈处理”,意思是,将递归函数调用化解为“一个由stack_push stack_pop stack_top等函数调用组成的循环式子”.这里 ...
- MIT scheme入门使用
在win7下可安装MIT-GUN scheme, 点开后有两个界面:一个交互式命令行界面:一个Edwin界面. 在命令行界面按Ctrl-G可以开始输入.在Edwin界面,输入完整命令后按Ctrl ...
- scheme Continuation
Continuation Pass Style在函数式编程(FP)中有一种被称为Continuation Passing Style(CPS)的风格.在这种风格的背后所蕴含的思想就是将处理中可变的一部 ...
- 递归——CPS(一)
程序中为什么需要栈stack? 普通的程序中,接触到子程序和函数的概念,很直观地,调用子程序时,会首先停止当前做的事情,转而执行被调用的子程序,等子程序执行完成后,再捡起之前挂起的程序,这有可能会使用 ...
- Scheme实现二叉查找树及基本操作(添加、删除、并、交)
表转化成平衡二叉树 其中有一种分治的思想. (define (list->tree elements) (define (partial-tree elts n) (if (= n 0) (co ...
- Scheme r5rs letrec的用法
说明,这是r5rs的用法. (letrec ((<variable> <init>) ...) <body>) 假设((<variable> <i ...
随机推荐
- kibana 统计每天注册数
- 【转】 树莓派学习笔记——I2C设备载入和速率设置
原文网址:http://blog.csdn.net/xukai871105/article/details/18234075 1.载入设备 方法1——临时载入设备 sudo modprobe -r i ...
- 使用WCF实现SOA面向服务编程—— 架构设计
原文地址:http://www.cnblogs.com/leslies2/archive/2011/03/29/1997889.html SOA本身就是一种面向企业级服务的系统架构,简单来说,SOA就 ...
- UVa 147 Dollars(硬币转换)
题目大意:给出五种硬币,价值分别为 1,5,10,25,50,.当给出一个价值时,求出能够组合的种数(每种硬币可以用无限次). 思路:完全背包, dp[i][j]表示总数 i 能够被表示的种数.状态转 ...
- 《Algorithms 4th Edition》读书笔记——2.4 优先队列(priority queue)-Ⅳ
2.4.4 堆的算法 我们用长度为 N + 1的私有数组pq[]来表示一个大小为N的堆,我们不会使用pq[0],堆元素放在pq[1]至pq[N]中.在排序算法中,我们只能通过私有辅助函数less()和 ...
- DBA 经典面试题(4)
1.如果信息采集管理系统(ICM)崩溃了怎么办? 答案:所有其他的管理器都会继续工作.ICM只会处理队列控制请求,意思是开启和关闭其他并发的管理器. 2.你如何加速打补丁的过程? 答案: ...
- Visual Studio 2015 使用ODP.net进行EF开发
刚转了新公司,以前公司都是用VS+MSSQL作为开发工具的 现在新公司由于数据库是Oracle,而且新公司比较小规模,开发团队也没有什么规范 访问数据库的方式一直使用ADO.net的DataTable ...
- Ubuntu12.04 Git 服务器详细配置
Git是一款免费.开源的分布式版本控制系统,用于敏捷高效地处理任何或小或大的项目,学过Linux的都知道,Git的优点我就不再多说了,我也是很喜欢Linux的.今天我们一起学习Git服务器在Ubunt ...
- 如何搞定前端资源服务跨域问题之nginx篇
问题描述 1.首先让我们先看一张图 2.从图中,我们可以很清楚的看到当http请求的站点访问https的资源的时候会报出“Cross-Origin”跨域的问题.为什么会出现这样的错误,这是因为涉及到“ ...
- JQuery 之事件中的 ----- hover 与 onmouseover 、onmouseout 联系
hover([over,]out) 一个模仿悬停事件(鼠标移动到一个对象上面及移出这个对象)的方法.这是一个自定义的方法,它为频繁使用的任务提供了一种“保持在其中”的状态. 当鼠标移动到一个匹配的元素 ...