Cpu is fast and disk is cheap. This is our mentality nowadays. But that can be wrong, and even it's true, it's still good to think about optimization.

here's a simple benchmark of some operations: (if it's not aligned, not totally my fault:)


So, how to optimize those computationally-cost operations?
1. Lookup Table is always a choice if you know the input range and the resolution.
2. Taylor (or other) expansion is also a good choice
    Example:
    Tan(x) can be replaced by (1/(0.5-x) - 1/(x + 0.5))/pi : (with small difference:)
     Use gnuplot:
     gnuplot> plot [-0.499:0.499] tan((x)*pi), (1/(0.5-x) - 1/(x + 0.5))/pi

http://discussions.pbs.org/viewtopic.pbs?t=68073

Have you ever consider to add new lines at some positions in VI?

when I editting HTML w/ vi, found out that one of the file all the new lines are gone, and so, I want to add new lines after
so I can see them clearly.

This is how I do it in vi:

:1,$ s/>/>^M/g

^M is to hold the ctrl-V and then press 'm'.

Good luck!

replace words in multiple files:

for i in `ls *.txt`; do vi -c ":1,$ s/abc/abcd/g" $i; done

But you have to close the file one by one. It might not be a bad thing because you might want to make sure the commands is correct anyway. The following method is the real "automation" if you can make sure your commands are correct.

for i in `ls *.txt`; do vi -e -s tmp_cmds $i; echo $i; done

and in the tmp_cmds, I put:

:1,$ s/abc/abcd/g:wq!

Another way is to do it with awk or curl:
counting:

watch

世界上最无聊的莫过于写blog这种事了。尤其是对一个天天和程序打交道的人来说。
好了，写些流水账吧。
今年有坏有好一整年，（废话，哪年不是啊？）
好多个开始：工作，飞行...
好多个结束：学业，在Santa Cruz的生活...
LiangZhuo同志说，只要有那么两个，我就说好多好多。。。。特别夸张。keke.... 是啊是啊，我也发现了。好吧，多举几个例子吧：开始玩facebook, 结束天天上去看好友都在做什么。开始吃免费饭，开始讨厌吃免费饭。
开始爱哭。开始爱笑。开始神经质~~~ hehe... (有这么说自己的吗？//kick bright)
结束写故事，开始写流水账。
不说了，饿了。

祝大家新年快乐。圣诞好！