Apple: the Blaze browser test was flawed

Yesterday’s report by Blaze.io claimed that the Android Browser is on average 50% faster than the iOS browser. Now Apple have this to say to Blaze: “Nuh-Uh”. Blaze posted an update, saying they are looking into the issue Apple pointed out.

It turns out that the embeddable web browser in iOS 4.3, the one used inside the custom app that performed the tests automatically, isn’t as good as the stand-alone Safari browser, missing several key performance updates…

Mobile Safari on iOS 4.3 received the Nitro JavaScript update, improved caching and asynchronous mode. Third party apps can embed a browser element, called UIWebView, which can be used to render web pages inside a third party app.

Some use it to offer custom web browsers for iOS, while Blaze used it in their app to load thousands of web pages automatically and measure the load time accurately (they used the same approach in Android). And this is the flaw in their tests, says Apple. UIWebView is not quite the same as Mobile Safari and is a worse performs.

Well, JavaScript didn’t affect load times much (the iOS 4.2, which lacks Nitro, show as much) but caching and asynchronous (that is multithreaded) rendering do have adverse effects in loading speed (as much as two, two and a half times The Register and some devs claim).

So, yesterday’s results should be taken with a big pinch of salt. But I can’t help but wonder – why gimp third party apps that try to render a page? I know that anything that competes with a native app is evil, but come on.

Source 1, Source 2