FameofLight    ACRush: Finally the god of programming arrives 

  

FameofLight    ACRush: do you follow a random practice , or you practice question by topic , in your early days 

ACRush    FameofLight: No, our team practiced 3 or 4 full contests a week. 

  

pt1989    ACRush: Can u tell us whom or what do u credit for ur success? 

ACRush    pt1989: I think more practice is the way. 

  

ktuan    ACRush: how can you arrange your study with 3-4 contests a week ? 

ACRush    ktuan: UVA, SGU , ZJU and PKU is enough. 

  

pt1989    ACRush: what is ur method of practising? random or organised? 

ACRush    pt1989: contest env.  We always solve programs in contest env. 

  

Sunny_05    ACRush: wat is ur daily routine ?? i mean hw do u practice everyday ?? 

ACRush    Sunny_05: We only have weekly routine for practice. 

  

binarywithme    ACRush: what u think about world final problems..level of difficulty 

ACRush    binarywithme: A little harder than 1000p. 

  

MB__    ACRush: Which of these do you do after you submit your solution on TC: test solution, re-read problem statement, check code line by line 

ACRush    MB__: test for at least 10 cases. 

  

martins256    ACRush: how many problems have you solved on OJ ? 

ACRush    martins256: about 2000. 

  

kcm1700    ACRush: How many problems have you proposed on OJ or anything like this competition or something...? 

ACRush    kcm1700: about 2000. 

  

vijay_comc    ACRush: Who is your Arch Rival in Programming ? ;) 

ACRush    vijay_comc: Petr, I think. 

  

SomethingWrong    ACRush: What is your favourite Online Judge? :) 

ACRush    SomethingWrong: SGU. 

vexorian    ACRush: Why SGU? 

ACRush    vexorian: More tricky testcases. 

  

Dee2306    ACRush: Being one of top programmers, do u believe its just practice which makes u perfect.. or some of u are born genius ?? :) 

ACRush    Dee2306: It's at least 80% due to practice 

  

stormsky    ACRush: how to practice ACM for a ACM beginner? 

ACRush    stormsky: practice in contest env. 

  

piva    ACRush: Do you train more of one category of problems? Or you just solve problems at random? 

ACRush    piva: I do more practice in the ones I have troubles with. 

  

pt1989    ACRush: what are ur interests other than programming? 

ACRush    pt1989: WarCraft. 

  

vijay_comc    ACRush: Petr is already married. No plans to compete him in that area ? :D 

ACRush    vijay_comc: ....not yet. 

  

Sarkin    ACRush: Is analyzing algorithms an essential part of learning algorithms? 

ACRush    Sarkin: Not all, I think coding is in practice room is the rest way. 

  

abdessamad    ACRush: can believe it! when i see your challengers :D :D 

ACRush    abdessamad: It's one thing that Petr did much better than me. 

  

stjepan    ACRush: where did/do you practice most and how? 

ACRush    stjepan: And join contest. 

  

binarywithme    ACRush: why u choose c++ as a default programming language 

ACRush    binarywithme: er.. Stl is the first reason. Another one may be efficient. 

  

geekru2    ACRush: during contest, what kind of problems do u enjoy doing? as in type of problem? 

ACRush    geekru2: dp and network-flow. 

  

reginachris    ACRush: What's the best OJ to start with for practice (UVa, SPOJ, TC, etc.)? 

ACRush    reginachris: USACO 

  

ferry    ACRush: What do you do when you can't solve a problem or you don't know what's wrong? 

ACRush    ferry: I will try to pass it in practice room. I always do that. 

  

geekru2    ACRush: Do you solve puzzles and mind bending questions to improve your problem solving techniques? 

ACRush    geekru2: no. I don't like such problems like suduko. 

  

Sarkin    ACRush: What algorithm types helped you in the IOI? 

ACRush    Sarkin: dp. 

  

abdessamad    ACRush: did you like chess? 

ACRush    abdessamad: yes. 

  

geekru2    ACRush: when in a team event(IOI etc), do you dominate while programming? 

ACRush    geekru2: yes. I guess. 

  

kcm1700    ACRush: how do you prepare data for the challenge phase? 

ACRush    kcm1700: The trickiest one, of course. 

  

binarywithme    ACRush: sir what is the best way to learn DP. 

ACRush    binarywithme: TC contest. 

  

[dasha]    ACRush: When u were beginning to compete,did u have any problems? Like low speed of sloving, or maybe some particular method u couldn't understend for a long time, sth else? If yes, how did u get over that? 

ACRush    [dasha]: USACO is a really good place. Especially for the beginners. 

  

frank44    ACRush: how long did the usaco practice take you? 

ACRush    frank44: half one year. 

  

binarywithme    ACRush: what u think math should b strong to become a Gud programmer 

ACRush    binarywithme: The mathematical foundation is helps. 

  

khanhptnk    ACRush: excuse me, do you think what we should prepare before an important contest ? 

ACRush    khanhptnk: relax, and rush some simple problems. And solve some problems in a contest env. is also helpful. 

  

Sarkin    ACRush: Do you recommned reading "Introduction to Algorithms"? if you know? 

ACRush    Sarkin: Really good. 

  

coder29    ACRush: which parts do u advice to read in "Introduction to Algorithms"? 

ACRush    coder29: All parts are perfect. except complexity. 

  

MB__    ACRush: do you do any sports but topcoding? 

ACRush    MB__: soccer. 

  

vijay_comc    ACRush: Complexity in CLRS is flawed ?? :o 

ACRush    vijay_comc: 7-8 

  

Sarkin    ACRush: What do you mean except comlexity? 

ACRush    Sarkin: That's only my opinion. 

  

pcaldeira    ACRush: which skill do you consider most important on TCHS/ioi-style contests? 

ACRush    pcaldeira: algorithm skills and coding ability. 

  

coder29    ACRush: I am newbie...which OJ is better fr me...USCO or UVA? 

ACRush    coder29: USACO 

  

hakami    ACRush: Are you useing library or typing from first for SRMs? 

ACRush    hakami: some includes and basic untilities. 

  

vrajesh1989    ACRush: Will you try to prove your algorithms to yourselves during contest time or just believe your intuition and start coding? 

ACRush    vrajesh1989: Yes, it's very important for me. 

  

Sarkin    ACRush: Do you use algorithm analyzis when solving a problem to see it's effeciency? 

ACRush    Sarkin: Yes. the effeciency sometimes is more important than correctness in TC contest. 

  

coder29    ACRush: basically i want 2 to do pratice on DP and greedy types...is TCs' room are sufficient..? 

ACRush    coder29: A bulk of dp tasks in SRMs. TC's room is good. 

  

rajeshsr    ACRush: In short, what is the strategy for becoming a good coder, according to u? 

ACRush    rajeshsr: Practice is the way. 

  

Lingmum    ACRush: How can we improve our speed of solving the problems,more problems? 

ACRush    Lingmum: More practice. 

  

stormsky    ACRush: do you often practice TC?and how to practice? 

ACRush    stormsky: Yes. But in fact, I prefer to doing past Regionals and finals. 

  

sahiltiwari    ACRush: how many hours you practise per day ? 

ACRush    sahiltiwari: 4-5 

  

MohammadReza    ACRush: What do you do to make your code becomes bugfree although the big size? 

ACRush    MohammadReza: test more tricky cases.(corner cases) 

  

rajeshsr    ACRush: Sorry to be repetitious, But I want to know what is the strategy of practice you employed? U select some random problems from OJs and solve or try to master a particular domain like DP by solving problems based on that or any other thing? 

ACRush    rajeshsr: set problems of past regionals and finals. 

  

puneetkp444    ACRush: Can you suggest some ways to improve problem solving abilities 

ACRush    puneetkp444: practice all kinds of problems. 

  

alft7    ACRush: what do you usually do before a very important competition, practise or something else? 

ACRush    alft7: practice until 3 days before it. 

  

ferry    ACRush: What is the hardest problem you've done? 

ACRush    ferry: I in WF2007. 

  


英文:o'reilly 

中文:博文视点 电子工业出版社

定价:45元,太贵,适合于从图书馆借来,每天睡觉前看一会儿。。。

 

下面是我的读书笔记,《短文标题,小标题》:

 

Beauty is in simplicity

1.阅读公认专家的编码

2.使用更短的方法 5-10行也不算极端

 

before your Refactor

1.起点是:清理已有的基础代码和基于这些代码写的测试

2.避开重写的诱惑(重新测试麻烦)

3.逐步增加小功能

4.新技术不是重构的理由

 

coding with reason

1.变量最小作用域

2.空行分割

3.需要内嵌的代码-》封装成函数

4.函数要短,古老的24行限制依然有用

5.函数参数四个以内(相关参数组合成局部对象)

6.封装且接口要窄

7.setter尽量少用

 

continuous learning

1.杂志、博客、微博

2.亲自编写

3.导师(最厉害的家伙)

4.看源码学习

5.午餐学习会、用户组、小范围讨论会(给别人讲授将会大大激发学习热情)

6.加入技术小组

7.参加讨论会,获得网上免费资料

8.使用播客,打发路上时间

9.《the pragmatic programmer》

 

do lot of deliberate practise

rt

 

未完待续。。。



今天恰巧错过CF,稍微用点时间扯扯我一直想写的文章《课程设计那点事》。

提笔的一瞬,突然不想扯废话了,那就我用我的“简洁”来叙述吧。

有了自信,就有了一切。————题记

从大一下C++说起。
//《无与伦比的美丽》
我一开始还是蛮有抱负的,重金买了C++primer,每天都看一些,每次下课之后都会围在老师严悍

身边问若干问题,可是,不知道哪一天,我就坚持不下来了,C++primer开始接受灰尘的洗礼,甚

至教材也不看了,课也翘了,上机作业也完不成了。
//《披着羊皮的狼》
于是,面对对象那一块我一点都不会了,甚至到现在我依然对虚基类等概念一无所知(现在是不想

知道),于是,我的C++就这么华丽丽地悲剧了。期末考试时,我觉得写个汽车类是世界上最难理

解的东西,考前刚背的都忘了。。。

然后,大二上开始了C++课程设计。

这中间还想插一段南理工校赛的事情,但是我发现我到现在还没有进入主题,罢了,那一段的结尾

是“决赛0题收场,又一次打击”。
//《倾国倾城》
C++课设两部分,一个人的做的那个是啥我都忘了,这里说出来也不怕大家笑话,我抄袭了材料系

08级的C++课程内容,因为面对对象的内容完全不会,于是就拼凑了几个人的程序和最终的报告(

丢人,连结题题告都是拼凑的),后来程序还是有不少问题,不过老师审核的时候竟然只发现了一

个问题,哈哈,水过!

两个人的部分和刘铁俊组队。那还真有不少东西可说。(正文本来应该从这里开始,我竟然写了这

么多废话,关键是,有几个人能坚持看到这里???烦不了了,本来就是写给自己看的,下面争取

简略一点吧)
//《酒干倘卖无》
首先是自己收集到很多资料,这意味着我又可以拼凑出一篇完整的课设了,不过后来还是自己写了

,这些资料成了我帮助很多其他同学完成课设的基础。
/*这里打断一下
上文注释里都是现在划过我脑海的歌,这个既有环境氛围,又可以计时哈
估计下面的文字没什么人会看了,我放弃对文章连贯性的要求,胡扯了
*/
//《那些事,那些人》
刘铁俊一直在帮助别人(mm)做课设,所以99%是我搞定的,他那1%是因为最后演示时我刚彻底唾

弃vs,用他的电脑演示的,哈哈!
//《拯救》
后来做了一个背单词的软件,很水,代码大概一百多行吧,就是读进来若干单词,然后随机显示,

没有技术含量,作为参加过ACM培训的娃,读输入还是很在行的,然后最后展示的时候怕没有亮点

,就鬼扯说用的是stl里面的vector存储,后来果然在老师的笔记上看到“亮点stl”的字样,再后来

就是在交源码之前把一个开的很大的数组改成vector。。。。
因为觉得这个一百多行的代码作为大课程设计很不靠谱,于是在演示前的一个下午找学长用mfc做

了一个,果然又是很快就搞定了,然后老师本子上另一个亮点变成“mfc”,哈哈,两个亮点都是

吹出来的。
后来,不少遇到困难的同学找我帮忙,我基本上就是把自己收集好的资料改编一下给他们,比如说

把一个电话簿改成药品管理系统,仔细看里面的”药品数量“用的变量还是phone呢,O(∩_∩)O哈

哈~
最后如愿得了优秀,因为骗老师骗得太深了,之后不少人来找我帮忙,一共大概做了六个人的课设

吧,都是粗制滥造,搞不得好还会被传给学弟学妹,贻害后人。。。(不是我的错哦)


这里应该是文章的前面一小部分,可是写了这么多,下面继续”争取少说废话“。

//《今天你要嫁给我》《我要的飞翔》《》《玄色风》
//忘了记歌名的,发现千千的随机播放列表是同一个序列,返回一下又找到了

这里省略好多,就总结一下“大二上学的《数据库》课程对我非常有用”,会了数据库可以开发好

多有用的东西了,不再垃圾程序了。

寒假学了php,开发了njust poj ranklist,可以看实时排名,每天00:00记录一次,形成排名曲线

,大家蛮喜欢的。不过后来放到了一个不稳定的服务器上,数据库都被黑掉之后了,大受打击,版

本号定格在v0.69,好几个新功能还没上线停止了开发,我想,明年软件课程设计1我会重拾的吧,只

有没找到更好有用题目。

插一句,我现在课设的原则是不做没有用的东西,比如这次java课设酝酿了好久,搬运工写了一半

废掉了,多线程ftp上传下载备好资料并写了一部分之后也废掉了,最终在突如其来的报选题截止时

刻,看到ghostplant报了“java c++ c#评测系统”,Teenager Studio 报了“njust online judge

分布式节点”,于是大脑发热,报了“online judge rating counting system”,这个还没有评估难

度的题目,因为之前一天srm499申小号想混入div1失败,看到gxxlovegxx在div2里面拿到1700+的

恐怖分数,很好奇rating算法。。。

后来就开始辛苦了,下面分点叙述:
1。中文关于rating的资料少之又少,搜了半天都是一样的东西,找资料困惑了我很久,比如

Volatility的初始值是多少?问了白衣神也不知道,后来让在ubuntu里尝试chrome,自动翻译了

tpo的论坛一个讨论,找到了答案
2。tpo rating算法里面有一个看不懂得符号,解释是:xxx is the inverse of the standard

normal function. 这个在线翻译也没能告诉我是神马,于是查老师,查课表,找了一个有米国半年

访问学者经历的老师,告诉我是正态分布,他不是研究概率统计的,不知道怎么算,给了个积分的

公式,这又难倒我了,还好,又是用google发现apache的一个开源项目写好了java的库函数,于

是,果断下载,学习API,调试,期间,为了保证正确率,把还没开始的《概率论与数理统计》翻

了一遍,总算找到验算的数据了,这时,发现之前写的很多代码设计有问题,大大重构了一番
//《寂寞沙洲冷》《雨中飘荡的回忆》
3。如何获得比赛成绩,一开始我没敢再课设题目之前写上njust,准备采用抓取页流的方式获得比赛

成绩,后来很幸运,Teenager Studio的两位大神开放了API,直接通过web服务获取比赛结果,方

便多了!
4。其实我获得API的时候距离截止日期已经不到一个星期了,熬了几夜,中间各种破事,耗费无数

真元,总算搞定了
5。今天演示,很成功,老师比较满意,同学们的评价也不错
6。昨天看面对对象的书,发现我还是根深蒂固的面对过程思维,代码虽然没有bug地实现了所有的

功能,但是代码质量太差,除了我估计谁都看不懂我的代码。今天早上总结了一下,发现了七处需

要重构的地方,等过一阵子闲下来了,就把重构当做训练自己面对对象思路吧!

END
突然通知我周四要讲动态规划报告,我还什么都没准备呢,于是写完这篇文章,暂时放下所有的

java和oo,先搞定报告再说。

 

写完之后发现cf已经结束了,我的ACM生涯走偏了吗???

roba的咆哮最后指出,项目经验,有木有?有木有?

可是,我想问自己,ACM正规比赛经验,有木有?有木有???


poj3844水题,检查同于相同的数量即可,注意一下余数为零的情况,然后ans=西格玛c(2,n)

最近一直在搞java课程设计,一直在用java的ide,于是就悲剧了。。。

不停地TLE,时不时还来几次RE,于是我求助好友,看到他们c++的ac代码,和我差不多一样的,都过了。。。

import java.util.*;

public class Main {

    public static void main(String[] args) {
        //Scanner cin = new Scanner(System.in);

        Scanner cin = new Scanner(new BufferedInputStream(System.in));

        int a = cin.nextInt();
        for (int cas = 0; cas < a; ++cas) {

            int mod = cin.nextInt();
            int num = cin.nextInt();

            int[] sum = new int[num];


            sum[0]=cin.nextInt()%mod;
            for (int i = 1; i < num; ++i) {
                sum[i]=(sum[i-1]+cin.nextInt())%mod;
            }
            int[] count =new int[mod];
            count[0]=1;

            int ans = 0;
            for (int i = 0; i < sum.length; i++) {
                ans+=count[sum[i]];
                count[sum[i]]++;
            }

            System.out.println(ans);
        }

    }
}

这代码改成c++肯定过,可惜java超时了。

哎,人品不好

 

这里转发一篇《给JAVA同学刷OJ的建议》吧,以飨后人:

 

1. 如果时间或是空间很紧张,不要相信任何系统库。ArrayList, LinkedList, Hashtable, HashMap之类看起来很值得信任很好听的名字效率都低得可怜,不仅如此,还异常地消耗内存。这不是-O2下的vector,即便是std::map的效率都比HashMap高得多。唯一可以信赖的是数组,效率还过得去。

这里说一个小技巧,比如需要实现一个int->int的映射,可以先开一个数组,直接用mod定址,这是非常效率高的。在出现冲突时,再扔到一个HashMap里。一般来说,控制得好的话,HashMap里的总数应该不会达到总共的10%,然后整个时间和空间效率就基本可以满意了,也节约代码。

2. 语言特性和频繁地申请对象使Java不可避免地在效率上远落后于C/C++,但抛开效率,很多在C里会犯的低级错误都不会在Java里犯了,而且运行时抛出的异常有助于迅速检查代码。当测试过几个数据之后,再根据逻辑读一遍代码,就基本是通过了。通读程序很重要,有利于发现测试不易检查出的bug。

3. 如果有一个IDE将大大提高开发的效率,对于C或是C++来说,这种提高很有限,但Java的话则会很多。

4. regex效率尚可,乃一大利器,理论上说,Regional中(只要是Linux系统的Online Judge也可以)C/C++也是可以使用<regex.h>,但功能比java.util.regex里的少得多。

5. Java没有complex,给计算几何代码编写带来一些困难,但实现一个也不是非常困难。Math的功能尚可