Java on Linux vs MS-Win XP
I'm was at Superquest all last week. Superquest is an event for training computer science teachers to teach computer science. It's where we get together to share best practices in computer science education, and is generously funded by an industry trade group, the "Software Association of Oregon".
Superquest was being hosted at Western Oregon University, a small university at Monmouth, Oregon (about 20 miles from Salem which is Oregon's capital city).
I'm taking the Advanced Placement teacher training class for teaching Advance Placement Java, the language chosen by the College Board in the United States for the Advanced Placement Computer Science/Programming curriculum for high school students.
We're using the BlueJ IDE, an open source IDE used largely by educators for teaching Java.
The computer lab we were using at Western Oregon University had Windows XP on all the machines. The first day of class, I used Win-XP for a few minutes--but I just couldn't stand it anymore . . . I whipped-out my PCLinuxOS CD I'd brought along with me, and proceeded to install Linux on the machine.
Western Oregon University has lots of bandwidth going into its computer labs, so I then quickly run Synaptic and install all the updates. Then I go to BlueJ's site, and install the IDE. In a few minutes, I've got a well functioning Linux Desktop in front of me. I even plug-in my usb flash drive, and copy over my favorite Desktop wallpaper.
I'm not into benchmarking, but upon launching BlueJ, I couldn't help notice that it seemed much faster than it was under MS-Win. Since the guy next to me was running BlueJ under Win-XP on an identical machine, a rough comparison was easy.
I asked him to shut-down and restart his Win-XP machine, and I do the same on mine. After restart, we launch BlueJ at the same time. The results? On the Win XP machine, the launch to BlueJ ready screen was nearly 5 seconds. On my Linux machine, the launch to ready time was 1 second! Furthur repeats produced the same result.
The BlueJ IDE is written in Java itself. And while this is hardly a scientific comparison, it is a side-by-side comparison on identical equipment under identical circumstances.
Why is launching and executing a Java program so much faster under Linux than Win-XP? I don't know. In my more paranoid moments, I wonder if Microsoft has deliberately crippled the performance of Java under Win-XP. Microsoft hates the popularity of Java. It's not their own dot net platform, and we all know that Microsoft wants to own and control everything.