如何跨領域研究以及困難

　　選擇進入IT領域後，不斷學習是必然的。不管是在自已的專業領域，或是跨領域知識，都是要學習的項目。那如何開始學習跨領域知識，洪教授有分享他的心得，如以下網誌：

台大資工 洪士灝&PAS Lab
http://hungsh-ntucsie.blogspot.tw/2013/12/blog-post_23.html

遞迴版 排列組合(permutations) using c#

在codeproject網站有介紹跟code。

http://www.codeproject.com/Articles/37215/Permutations-in-C-Using-Recursion

寫程式這件事在學校與在業界中有何差異？

What's difference between college-level and corporate programming?


從台大洪教授fb分享的連結看到這篇，我想把他翻成中文，希望自已進入職場前多了解一些，甚至在研究所就開始加入一些觀點，來快速適應職場環境。也希望中文版的內容可以幫助到更多人。


(待)




quote from:
http://arstechnica.com/information-technology/2013/12/whats-the-difference-between-college-level-and-corporate-programming/
by Stack Exchange

About the emguCV running in Windows Phone 8

在十月份的時候我在EmguCV的社群發了一封留言，請教EmguCV有沒有支援在WP8 APP。就在11/3我收到回覆，內容為：

Sorry, Windows Phone 8 is not supported. Windows 8 App store refers to the Windows 8 store running on x86, not Windows phone 8 store running arm. Microsoft is having too many different platforms at the moment:

1. Windows Desktop (regular windows executable)
2. Windows 8 App store App running on x86 (used to be called the metro interface)
3. Windows RT App store running on ARM (e.g. MS Surface, Surface 2) 
4. Windows Phone 8 App store running on ARM (e.g. Nokia phones)

As silly as it seems, app build for one platform cannot run on the others. None of platform 2, 3 and 4 seems to be a success at the moment and market share is small. Emgu CV currently support platform 1 & 2. It is possible to support platform 3 but we do not see enough demand for it. We don't have a road map for platform 4 yet because open cv do not support Windows Phone 8. 

Best regards,

Canming



整理：
    EmguCV目前支援在x86架構上，所以Windows Desktop跟Windows 8 App可以呼叫來用。但是WinRT跟WP8是ARM架構，所以不支援。
    在最後一段他寫了一些原因，也似乎婊了一下Microsoft。Win8 App、WinRT App跟WP8 App都是你家的東西，但build出來的App卻不能全部支援，有點愚蠢，而且這三個東西現在似乎還沒成功，市場也小。
    WinRT目前需求不大，不支援。OpenCV不支援WP8，所以EmguCV也無法支援。

How To Become A Hacker by Eric Steven Raymond

What Is a Hacker?
    The basic difference is this: hackers build things, crackers break them.


The Hacker Attitude
 
    To follow the path:
    look to the master,
    follow the master,
    walk with the master,
    see through the master,
    become the master.

  So, if you want to be a hacker, repeat the following things until you believe them:

    1. The world is full of fascinating problems waiting to be solved.

    2. No problem should ever have to be solved twice.

    3. Boredom and drudgery are evil.

    4. Freedom is good.

    5. Attitude is no substitute for competence.

Basic Hacking Skills

    1. Learn how to program.

    2. Get one of the open-source Unixes and learn to use and run it.

    3. Learn how to use the World Wide Web and write HTML.

    4. If you don't have functional English, learn it.


Status in the Hacker Culture

    1. Write open-source software.
    2. Help test and debug open-source software.
    3. Publish useful information.
    4. Help keep the infrastructure working.
    5. Serve the hacker culture itself.



Reference:
How To Become A Hacker, Eric Steven Raymond,  http://www.catb.org/esr/faqs/hacker-howto.html#why_this

Assignment1_A

Assignment1_A 不用寫到任何CODE，算是簡單的。

My First iOS app

終於開始自學iOS app撰寫了！

教材是Stanford University的iphone development 開放式課程CS193P 。

買了兩本工具書，精通OBJECTIVE-C程式設計以及探索IOS5程式開發實戰。

新的語言一定要Hello World! 一下。

//ViewController.h
@interface ViewController : UIViewController{
    IBOutlet UISlider *slider;
    IBOutlet UILabel *textValue;
    IBOutlet UILabel *hello;
}

@property (strong, nonatomic) IBOutlet UISlider *slider;
@property (strong, nonatomic) IBOutlet UILabel *textValue;
@property(strong,nonatomic) IBOutlet UILabel *hello;

- (IBAction)changTextValue:(UISlider *)sender;

//ViewController.m
@synthesize slider, textValue, hello;

- (IBAction)changTextValue:(UISlider *)sender {
    int SliderValue=slider.value;
    textValue.text=[NSString stringWithFormat:@"%d",SliderValue];
    hello.font=[UIFont systemFontOfSize:SliderValue];

Parallel.Foreach取得目前元素之索引位置

想法：在影像處理run pixel上用了Parallel Foreach跟For，比非Parallel的快非常多，但是卻少了Index。在做pixel操作的時候有時需要知道要寫回哪裡。
實作：
            byte[] byteArray = binary.Bytes; //size:307200
            var query = byteArray.Select((val, idx) => new { Index = idx, Value = val });
            int[] result = new int[byteArray.Length];

            //平行 擇一上註解
            Parallel.ForEach(query, b=>
            {  
                if (b.Value >= 255)
                {
                    result[b.Index] = -1;
                }

            } //close lambda expression
                 ); //close method invocation

            //非平行 擇一上註解
            foreach (var b in query)
            {
                if (b.Value >= 255) result[b.Index] = -1;

            }



結論：驚!! Parallel方法花了25ms。非Parallel花了17ms。Parallel for是否又比Parallel foreach?!
             等國科會結束我再來比較一下Parallel for!!!

在WPF將RGB轉HSV調亮度後轉回RGB再SHOW圖 using Emgu CV

想法：因為系統需要由使用者動態調整亮度。
實作：
Code：
        private void slider1_ValueChanged(object sender, RoutedPropertyChangedEventArgs<double> e)
        {  //sourceBgr&sourceHsv 事先建立
            Image<Rgb, byte> rgb = sourceBgr.Clone();
            Image<Hsv, byte> Hsv = sourceHsv.Clone();
            CvInvoke.cvCvtColor(rgb, Hsv, Emgu.CV.CvEnum.COLOR_CONVERSION.CV_RGB2HSV);     
            int  newValue=(int)e.NewValue; //取得slider事件發生後的值          
            string s = newValue.ToString();
            label1.Content = s;
            if (Hsv != null)
            {
                for (int j = 0; j < sourceHsv.Height; j++)
                {
                    for (int i = 0; i < sourceHsv.Width; i++)
                    {
                        byte value=0;
                        if (Hsv.Data[j, i, 2] + newValue > 255) value = 255;
                        else if (Hsv.Data[j, i, 2] + newValue < 0) value = 0;
                        else value = (byte)(Hsv.Data[j, i, 2] + newValue);

                        Hsv.Data[j, i, 2] = value;
                    }
                }
            }
            CvInvoke.cvCvtColor(Hsv, rgb, Emgu.CV.CvEnum.COLOR_CONVERSION.CV_HSV2RGB);

            image1.Source = rgb.ToBitmapSource();
        }

紀錄：Hsv的v(value)是亮度，只要調整它即可調整該PIXEL的亮度。

在WPF做倒數計時器

想法：因為系統需要倒數計時器來執行拍照。
實作：

step 1: 加入System.Windows.Forms參考，並引入。
step 2: coding!
        private void button1_Click(object sender, RoutedEventArgs e)
        {
            timer1 = new Timer();//global
            timer1.Interval = 1000;
            second = 5;//global
            timer1.Tick+=new EventHandler(timer1_Tick);
            timer1.Start();
        }

        private void timer1_Tick(object sender, EventArgs e)
        {
            if (second > 0)
            {
                label1.Content = second ;
                second--;
            }
            else
            {
                timer1.Stop();
                label1.Content = "";
                BitmapImage bi = new BitmapImage(new Uri("/Image/p3.jpg", UriKind.RelativeOrAbsolute));
                image1.Source = bi;

            }

        }

Label用FontSize調字型大小。

EMGU CV的GPU測試運作

想法：目前系統用到emgu cv non-gpu的函式，是否改用gpu 函式效能就會提升？
實作：先run起emgu cv gpu的函式，以及針對靜態影像做簡易的影像處理。

Step1: GPU package

GPU for image processing is only available for Emgu CV rev 2.2.1 and later. Only package containing -gpu in its name (e.g. libemgucv-xxx-gpu-xxx) has GPU processing enabled.

• Install the latest cuda graphic card driver from NVIDIA on your running platform.
• Copy the cuda and npp dll files: cudart{bit}_{maj_rev}_{min_rev}.dll and npp{bit}_{maj_rev}_{min_rev}.dll to the execution directory
• Add Emgu.CV.GPU.dll to References
• Optionally put the following lines in the top of your code to include the Emgu.CV.GPU namespace.

Step2:
針對同一張800*533影像做灰階->FastCorner->轉bitmapsouce後用元件show圖，過程皆用Image類型運算。GPU cost time 是 82ms，non-gpu cost time是 17ms。


記錄：第一台ASUS BM6650 i5-2500 3.3GHz intel內顯 品牌電腦。執行時會有unknown error，初估是非nvidia的顯卡。第二台配備nvidia quadro工作站顯卡，成功執行。

結論：gpu函式再初始階段較耗時，在簡易的影像處理函式中無法展現高效能特性，或許在更高計算量的影像處理需要時，能展現其高效能之特性，像是KinectFusion(real-time 3D reconstruction)。