It’s Not About The Rendering Engine Anymore

Today, RIM launched their new 6.0 OS (and the new BlackBerry Torch) and by doing so, added its name to the ever growing list of products using WebKit as the basis for their Web browser. RIM’s move isn’t really all that interesting: they had no good browser anyway and it’s not as if they are first to do this.

The interesting discussion in my opinion is what Microsoft should do. Back in the Netscape / IE war, Microsoft was convinced that Web browsers would ultimately replace desktops apps (they weren’t wrong) and they invested a lot of money in creating Internet Explorer. That war is still on, although nowadays the players have changed. With IE, we’re now talking about Firefox and Chrome as the main competitors.

But what really changed however, is that the HTML/CSS rendering engine shouldn’t matter that much. We went from a 1 browser = 1 engine model to a model where we now have a couple of really good open engines powering many browsers.

Controlling the world’s browser market is one thing, but as Google is proving, that doesn’t mean you need to have your own engine. Chrome is built on top of Web kit, just like Safari is and countless mobile web browsers are. These products are still competing with each others and are still different from each other yet they are built on the same foundation.

What that means of course, is easier Web development and less browser-specific bugs. This is why I think Microsoft needs to stop developing its own engine and start using either WebKit or Gecko. Both of these are well done, support many web standards and are fairly easy to integrate in a product.

Of course, such a change can’t happen overnight and it’s a difficult thing for Microsoft to do, but I do believe that in the long term, it’d make the Web a lot better. Trident, Microsoft’s rendering engine since IE 4 isn’t exactly renowned for its spectacular support for standards. And really, I’d much rather see Microsoft invest 18 months developing Internet Explorer proper rather than waste 80% of that time coding the engine.