Steven Yong's Weblog

Although it's been a good start in last week as my favorite soccer player Michael Owen scored his first goal after returning from La Liga (Spanish League), I was in some kind of dull mood for the entire week.
I couldn't get anything done for the week, or, more precise, I didn't want to get things done. Officially this is my final week in Bangalore, India and I'm flying back next week. The environment here is peaceful, the weather here is fantastic, the people here is friendly and the food here is delicious... But, I miss my home country, Malaysia.

"If one has no friend around, how far he or she can go?"

- Steven Yong

I believe most Java programmers know that they should try their best not to call overridable methods in constructor, which by doing so most likely will bring debugging nightmare to the programmer who extend their classes. This rule is especially true if the class is purposely designed to be extendable. Just remember that the superclass constructor will be called before calling your own constructor therefore some variables probably don't get initialized properly when they are in the override method accessed by the superclass constructor. Fortunately, there is a good tool which can help you to detect this mistake if you can't be bothered to find it out. Although PMD doesn't stand for anything, personally I prefer to remember this tool with theits acronym "Programming Mistake Detector".

Ansari, an Iranian-American reportedly paid $20 million to become the fourth private astronaut and achieve her dream of becoming the the world's first paying female space tourist.

Once a friend of mine asked me why is it so inconsistent that the index of first element of Java array is 0 while the first index of java.sql.PreparedStatement's setters is 1. For example, see below code:

char[] alphabet = new char[26];
alphabet[0] = 'a'; // assign the character 'a' to array alphabet

PreparedStatement pstmt = con.prepareStatement("UPDATE EMPLOYEES
SET SALARY = ? WHERE ID = ?");
pstmt.setBigDecimal(1, 1800.00); // assign 1800.00 to the first placeholder '?' 
pstmt.setInt(2, 110592);

Actually the answer is pretty simple, which I just found out. It's all about the purpose of the design. As we know, every element in an array has the same size, so, it is convenient to consider the index of the first element of the array to be zero (just as in C and Java). Have zero be the index of the first element, the compiler can simply calculate the memory address of the element using this formula:

element address = array base address + (index X element size)

Although it is possible to use other values for the first index, it complicates the address computation. Consider following formula if we use 1 as the first index:

element address = array base address + ((index - 1) X element size)

Obviously, less CPU instructions will be executed if the index 0 of first element is used. In contrast, java.sql.PreparedStatement need not worry about this performance issue, the parameter setter methods are designed for human and they will be no different with using 1 or 0 to be the index of first parameter in terms of performance.

After watching the cool demo video of Xgl, an X server architecture layered on top of OpenGL, this person doubts Vista's UI is better. Larry explained that cool user experience doesn't necessarily turn out to be good user experience, whereby sometimes cool UX can reduce the usability of the system tremendously. And that's why some cool features (like floppy windows) are dropped before final release. I wonder among those cool desktop windowing system like Xgl, Apple Quartz Extreme, Vista Aero and even Sun's Project Looking Glass, which one is moving to the right direction of usability.

From NCSA Mosaic to Netscape Navigator, from Netscape Communicator to Internet Explorer, from Internet Explorer to Mozilla FireFox, and the web browser wars continues. In next round, we will see FireFox 2.0 vs. Internet Explorer 7.

As announced, Microsoft launched Zune yesterday. It really looks nice and feature rich (WiFi, 30GB of HDD, built-in FM, a 3-inch screen and the basic music, pictures and video playback). However, as a iPod owner, I'd say no surprise and I believe it's going to be just another iPod killer.

It took me hours to fix a bug which considers totally new to me and it's about the IE save/open dialog. On the report export page, I provide xls and pdf format of report for user to download or open directly. Somehow IE keep complaining file not found in temporary folder if I chose open though the file's correctly saved if I chose save. Eventually I found that it's something to do with the header of the HTTP response, wherein some browsers may fail when both Content-Disposition and Cache-Control are set. The problem is subtle and I couldn't find a clue from HTTP spec yet and somehow it works fine on FireFox with both additional features set.