Friday, July 09, 2010

Reviewing the latest speed wars in Web Browsers

Google raised quite a stir when they claimed to have the fastest Web browser, and Apple did the same when they claimed that Safari is the fastest Web browser. Clearly we can't have TWO "fastest Web Browsers". Where do things stand today?

Well, what is good about all of this stuff is that browser developers are taking a closer look at resource utilization of their browsers. For many years, as Web browsers added more and more features, they became more and more bloated.

I've been around a long time, and I can vividly remember when GNU Emacs was soundly criticized for being so large, using up so much memory. Well, those complaints went away a long time ago. Before we even had 1 GHz desktop computer systems, I found that GNU Emacs would load in just a few seconds on computers with as little as 200 MHz and 32 MB memory. It's so much better today. The typical Web browser is four or five times larger just to download, and the amount of virtual memory they use can easily exceed a factor of ten beyond what an even loaded Emacs would consume.

Where does that leave the typical Web browser then? Two years ago, Web browsers were probably near their all time low in terms of efficiency and performance. Yes, they were offering more and more features, but the cost of those features were becoming prohibitive.

When Apple Safari came on the scene, they took the browsing engine, KHTML, from the Konqueror file and Web browser, part of the KDE project. They released their improvements as a Web rendering technology called Webkit.

A little over two years ago, Google took that technology and used it to form the basis of their browser project. They called the browser Chrome and the source code project Chromium. They have made numerous performance improvements on top of what Apple did, and the results were staggering.

When Google made Chrome available, it shocked many people with how much faster it worked, especially on Google sites. A few years ago, that technology was immature. Other features that people commonly used were noticeably missing. Many of those features are now included, but Chrome still tries to maintain a modest appearance. It's worked great, and Chrome moved into the number three spot behind Internet Explorer and Firefox in several usage reports.

This competition has proven to be very helpful. Microsoft has done a lot with Internet Explorer to improve its reliability, performance, security, and overall usefulness. It is so much better in Versions 8 and 9 that it is hard to fathom why so many people continue to even use Versions 6 (from XP) and 7 (from Vista).

Mozilla, which has lost some ground to Chrome, has been vigorously working on improving its performance on several fronts. Mozilla had already been working on a new generation of browser capability so that it would be technology that would be portable to various hand held devices. They've taken a lot of what they've learned and applied it to the Firefox project. Version 3.6.6, the current version, is already improved a lot, but just wait until you see Version 4.0. Currently called "Minefield" because it is undergoing nightly build testing, it is nevertheless in its second test release, and we ought to see it released later this year. I've tested it a number of times and it looks really good.

I already wrote the other day about another project that the Mozilla Labs has been producing, the Mozilla Prism project, to provide Web applications. That work is also bringing new ideas to the table, and it has been effective.

Opera, which has been around a very long time, was once known as the fastest browser, but it had become large and sloppy, much like the others. Version 10.60, recently released, has many positive changes to reverse that trend. Like the other browsers, it is much faster than its previous version.

It looks like what Google has done to awaken the industry has been a positive thing. It's worth giving Google Chrome a try, but it's also worth investigating some of the test versions of the new browsers. If you can spare the time and periodically test a few of them, and provide feedback and defect reports, it definitely helps to improve the browser landscape. I try to do so at least periodically.

No comments: