Real Programming

2008-07-22T15:18:00.002-07:00

Programming languages have come a long way. Managed programming environments offer a lot of great benefits; one of my favorites is dirt-easy event handling. Underlying coding environments (Java run-time environment, Perl, .NET) have made cross-platform coding far too easy. So the world of programmers keep developing larger development environments, more human-language-like syntax, and keep patting themselves on the back for jobs well done.

But it gets me thinking.

How much of today's programming is wasteful? Here's something I learned about assembly language today. If you want to zero the value of a register, you could use the statement MOV EAX,0, and that would move the value of 0 to the EAX register. In total this command uses 6 bytes. However, now I realize why I never saw that command. Instead I saw XOR EAX. Why? XORing any value to itself results in the value of 0. And guess what! It only takes 3 bytes. That's HALF the code!

Here's another thought from Joel Spolsky. I will use my own words.

The structure of a standard C null-terminated string is thus: "Hello, world!", 0
You see that the last character is the value 0, or NULL. Easy enough. Now, how do you determine the length of this string? Using a loop structure, you check each character one-by-one to see if its value is NULL, and you keep track of how many times you loop. For our example, "Hello, world!", the processor would loop 13 times, and then, on the 14th loop, would find the NULL value. So it takes 14 loops to determine the length of 13-character string. Now, how many loops would it take to find the length of a 756-character string? See where I'm going? The larger the string, the longer it takes just to find its length.

Now, Joel's thought was this: why not put the length of the string right before the characters? So "Hello, world!", 0 would now look like this: 13, "Hello, world!". How many loops would it take to find out the length of this string? Zero. Why? All you have to do is look at the first byte (or two, or four, depending on how you chose to do it) of the string. So even with a 756-character string the program only need look at the first number.

Sure, every time you update the string you have to update the length, but all-in-all, which version is more efficient? I may do some assembly research on my other blog, http://lowlevelprogramming.blogspot.com/.

The Best Software Writing I

2007-03-16T11:55:00.001-08:00

I created this blog to share my thoughts on computer technology. Sometimes I have odd philosophic thoughts about technology and its connection with people, and sometimes I just like to debate the best way to implement new ideas. Network security, freedom of exchange, responsible computer use.

Yesterday, I began reading a book titled The Best Software Writing I. The sections were selected and introduced by Joel Spolsky. I've read four sections of it and already feel as if my abilities with technology are mediocre. Will I ever think like these people? I hope so. The book shares the viewpoints of several authors on different subjects in technology. (On a side note, it was the section titled How Many Microsoft Employees Does It Take to Change a Lightbulb? that got me to buy the book.) This book motivated me to start this blog. Thank you, Joel.

Kenny Thinks

Real Programming

The Best Software Writing I