Lifehacker pointed to a list of software that claimed to be lightweight apps that replace certain memory hogs. The last two (original) articles are a great read, however the lightweight apps list is pretty poor.

The problem is several fold. His article is basically a rehash of the articles on The PC Spy listing the bests of specific categories, however, the original article wasn’t really about lightweight applications. It was about the programs that slow down Windows the most. There were a lot of complaints about items like Firefox (which will cache all the pages it visits until it runs out of memory) and Trillian (by no means lightweight, but an great application). Lifehacker escalated the problem by listing them not only as lightweight apps, but as open source substitutes. Not all of the programs he listed are open source and ’substitutes’ for what? Some you can assume are substitutes because there are dominate programs in that field, i.e. Foxit Reader as a substitute for Adobe. Others aren’t really substitutes at all. For example, uTorrent is the lightweight substitute for torrent downloads instead of… Sure there are popular bit torrent programs out there, but none of them hold a dominate or ‘default’ position that you would find substitutes for.

This got me thinking though. What exactly is a lightweight application and how would you measure that? In my opinion a lightweight application is something that I can quickly open to perform a quick task. The most important factor in a lightweight application then is launch time. Next to that would come factors like limited memory usage and low footprint (i.e. doesn’t try and take over my desktop, doesn’t take up much hard drive space, doesn’t pop anything up when I boot up my computer, etc.)

I wrote up a quick little c# program that launches other applications. I measure the memory usage, peak memory usage, and processing time. I’m not exactly happy with it or rather I don’t trust it right now, but I should probably try it with non-web based applications. Because I am not happy with it I spent a lot of time procrastinating talking about it. I decided I would finally finish one of my half done blog posts and this one got picked.

I decided to see what the lightest weight browser was. I used the code to take the average launching time of each of three browsers (IE 7, Opera 9.1, and Firefox 2) 10 times with three different start up options (Blank page, MSN.com, Google.com). Each browser was freshly installed and configured with no bookmarks, clean cache, no history, and set to launch a blank page by default. I ran it on a Athalon 3200+ with a gig of ram and XP home SP2.

The reason I chose this test was that I wanted to see which was the best ‘default’ browser. A major factor for the best default browser is how fast it opens. Currently I have IE7, which I never use, Firefox, which is my default but rarely gets used, and Opera, which I use all the time. The reason firefox is my default is because if I click on a link and Opera isn’t open, it is likely I just want to check out real quick, I want this to be as fast as possible and I feel Firefox does that.

Here are the results:

It’s the average of 10 runs, but I think the network causes too much distortion in the msn/google data. It doesn’t seem right that the simple google page takes longer to load than the busy busy msn page. I blame the network for this. I also can’t explain why the local blank page takes longer to load than msn in the case of Opera and Firefox. Caching maybe? Not just page caching but process caching. Not sure. However, if we just compare the local blank page between the three programs we have IE 7 as the fastest, followed by Opera, then Firefox. I am not suprised that IE is the fastest, but I was suprised that my clean, extensionless firefox install took longer to start than Opera.

So will I switch my default browser? No, I don’t trust or like IE and even if the data is correct and even if Opera is slightly faster than Firefox, Opera is set to continue browsing from where I closed it (like Session Saver for Firefox). So in the end Firefox is still the best choice for me to quickly open up a webpage.

I will probably try this on some other set of applications if for no other reason than to find out if my program really works and if the huge variance I saw in the browsers was due to the network.

This entry was posted on Friday, February 16th, 2007 at 8:13 pm and is filed under Software, programming, projects. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.