Google’s Redefinition of the Browser As Platform

According to StatCounter — one of several sites that keep statistics on Internet usage — Google Chrome became the most widely used browser in June on a global basis. Basing its information on tracking more than 15 billion page views, Chrome surged past Firefox late last year and overtook Microsoft’s Internet Explorer in the May-June time frame.

The principal reason for this success, I believe, remains the aspect that initially distinguished Chrome when it was released: its speed. It loads pages visibly faster than all other browsers. Part of this speed is attributable to Google’s excellent V8 JavaScript engine, which converts JavaScript into native machine code. The company’s relentless pursuit of speed, however, does not stop there. For example, Chrome was the first major browser to render PDF files directly in the browser, rather than calling a plugin that loaded the Adobe Acrobat reader. As a result, PDF documents sprang to life in Chrome, while they seemed to trudge along slowly and inexorably in competing products.

If Google had just left Chrome be as a speed demon, it would surely have competed well with the other browsers. But instead, Google soon realized that by adding new features to the browser, it could do what few other companies had done — provide both content and the delivery mechanisms for the Web experience. (For reference, google.com is the #1 site worldwide for Web traffic.) In this sense, they took a page from Apple, which has taught the computer industry that if you control most of the key points in the content pipeline, you can create a unique user experience. And excellent user experience breeds a deep loyalty. (This loyalty is visible in statistics from other Internet monitoring services. While not all place Chrome in first place, as they do not all measure global statistics, they consistently show Internet Explorer usage dropping during weekends and holidays and Chrome usage increasing significantly during the same period. This would imply that IE is used in business due to IT policies, but when users get back to their own machines, they prefer Chrome.)

In pushing beyond raw speed, Google has articulated a sweeping vision of the browser as a platform. And to back that vision, it has undertaken several impressive initiatives that will surely separate Chrome from the rest of its competitors.

The first of these is a deep exploitation of HMTL5. While all modern browsers boast of their commitment to HTML5 (I’ve discussed this regarding Internet Explorer 9), few have delineated so long a list of commitments. The most notable of these is the port of the company’s Google Gears (subsequently renamed just Gears), which was one of the first technologies to enable browsers to work with Web applications even when offline.

A second initiative has been to expand the programmability of Web pages with a new language, called Dart, which is a simpler way of coding JavaScript. It competes with CoffeeScript (among other languages), which seems to have a head start in terms of popularity. Behind Dart’s development is Google’s experience with the remarkable Google Web Toolkit (GWT), which translates Java into JavaScript accurately and reliably. (Even years later, GWT still strikes me as a remarkable and under-appreciated feat.)

Google is now taking programmability a step further: compiling C and C++ into JavaScript for running in Chrome. At this year’s just-completed Google I/O conference, I sat in on several sessions presenting the technology, commonly referred to as NaCl (pronounced “nackle”), which is an acronym for “native client.” I watched a demo of an OpenGL app written in C++ be slightly modified with #ifdef statements, then compiled and run in Chrome using NaCl. The process was straightforward, although from some glitches and comments by the presenters, it was clear that the project is still on the bleeding edge and not ready for more than experimentation. (Despite this, Google showed several apps and games that had been ported to Chrome via Nacl.)

I attended other sessions that explored support for audio in Chrome as well as building business apps for Chrome. The picture presented, and made entirely real in those presentations, was that Google views Chrome as a major platform. It does not see the browser as a mere tool adorned with plugins that add small slices of functionality, but a universal client-side target for development of all kinds in all major languages. To this end, Google is also moving towards what it calls “Chrome Packaged Apps,” which are apps that are downloaded from the Google Store and installed via the browser, primarily for use in the browser. The packaged apps also run on Android, and most platforms outside of the browser and can run offline.

The distinction between regular apps and browser apps is getting increasingly difficult to define. If you find this direction, as well as compiling and running C++ apps in the browser, as interesting as I do, Google Chrome is the browser of choice to participate in the revolution.


Small note: The traditional ways of writing extensions for Chrome still work. We recently ran an article on writing Chrome extensions). I believe you’ll find it surprisingly easy.

— Andrew Binstock
Editor in Chief
alb@drdobbs.com
Twitter: platypusguy