To say a simple thing

Saturday, December 31, 2005

dusk of cuckoo (杜鵑鳥的黃昏)

杜鵑鳥的黃昏 - 作詞:莫凡 作曲:莫凡

你說 秋天走了你也要走 悄悄飛向天的盡頭
你已習慣自由的生活 和遼闊的天空
你說 宿命總是要你遠走 注定一生落落的寂寞
讓我今生為你守候 作你最溫柔的風

誰知闖入紅塵 一旦醺然入夢
無聲悲泣哀鳴 任憑風吹雨落
但願重上枝頭 寧作花間粉紅
好夢一場總成空 化作縷縷相思游

是否 只是一個難圓的夢 蒼天靜靜向我訴說
終於知道你是那個 永不過夜的寒冬

註:原來杜鵑鳥是布穀鳥!
  我家的咕咕鐘啊!布穀!布穀!布穀!

broken hearts (受傷的心)

受傷的心 - 作詞:吳宗憲 作曲:吳宗憲

受傷的心已經無閒 無閒聽你擱講分明
就欲離開的行李內 找無你我的將來
男人的心嘛真堅定 擱再講感覺無著你的愛情
這張無留住址的批 代表阮永遠無後悔

啊... 傷心傷心的話 請你放置你的心肝底
無論天涯 亦是海角 阮永遠欲走乎你找
啊... 愛情愛情是假 何必聲聲句句欲挽回
若是日頭從西邊出來 阮還是決定欲走乎你找

Monday, December 12, 2005

Toward freedom

Some days ago, I decided to get rid of Matlab because
  1. It is heavy. It takes at least 3 seconds to start itself and usually the programs are running slowly compared to the programs written in Python.
  2. It is propritary. Although many people in the control field use Matlab to do simulations, I don't agree with that. Why not use open source software instead? For operations with matrices and transfer functions, I highly recommand Scilab. As for numerical calucations, I think any programming language will perform better than the Mablab programming language.
I use the Python programming language to do numerical calculations, but it don't have a built-in library to plot figures. So at first I use Python to generate the data needed for plotting and then use Matlab to load those data to plot the figures. Few days ago, I try gnuplot. Although I've known gnuplot some years ago, I do not give it a try. To make gnuplot work with Python is not hard. The following is my solution (not elegant but works)
def gnuplot(d_name):
gnuplot_cmd = "splot '%s' using 1:2\n pause -1\n'" % d_name
fcmd = open("XXX.dem", "w")
fcmd.write(gnuplot_cmd)
fcmd.close()
system("XXX.dem")
Today, I decided to get rid of IE because it does not conform with the W3C standard. Now my main browser is Firefox.

世代傳承的意志 時代的變遷
人們的夢 只要人們繼續追求自由的答案
這一切的一切都將永不停止

這世界 沒錯 一個追求自由任憑選擇的世界
就在每個人的眼前無限地延伸
如果我們的夢想可以引導你的方向的話 就去追尋吧
在名為信念的旗幟之下

Friday, December 02, 2005

Convexity

Recently I enjoyed surfing homepages of famous people in the control field. At first, I don't even know those people have a homepage. I came across their homepages because I wanted to find an old paper whose electronical version was not available through school's library. So I used google to search the paper's information. And found some authors place pdf files of the papers in their homepage, which is a nice idea from my point of view. I also found there are many interesting things in their homepages, e.g. their lecture notes, old papers, books, biography...

I especially like to see their lecture notes, because they are usually written in a compact way, only the main and important concepts are listed in the notes and tedious calculation are omitted. This helps me to quickly catch the most important idea that the authors want to convey.

Today I read three different people's lecture notes:

  1. Stephen P. Boyd: Convex Optimization in Electrical Engineering.
  2. Carsten W. Scherer: LMI Relaxations in Robust Control.
  3. M. C. de Oliveira: Background material on convexity.

I found one thing in common: they all have the following quote

The greatest watershed in optimiztion is not between linearity and nonlinearity, but between convexity and nonconvexity.

Rockafellar