Apple has admitted that the embedded Web viewer used to execute Web applications saved to the iOS home screen does not include optimizations made to the Safari with the iOS 4.3 release. This also affects native apps using the platform's UIWebView API.
The admission by Apple means that conspiracy talk can be halted. Since Web apps are a popular way "around" rejections from the App Store (e.g. the official Google Voice app before Apple relented, Readability), some developers theorized that Apple was slowing Web apps on purpose.
This also means that the browser performance tests run by Blaze Software and released on Wednesday were in fact flawed. The site, which offers a free service for measuring mobile web performance, ran a test comparing Android and iOS browsing performance. However it used custom apps for both Android and iOS, and the iOS version used theUIWebView API, which Apple said was lacking in the performance improvements in Safari in iOS 4.3.
The improvements in iOS 4.3 for Safari include the Nitro Javascript engine as well as "certain Safari caches and the browser's 'asynchronous' rendering mode."
Flawed or not, the statements by Apple admitting the differences between the two are likely a response to the report, so that is a good thing. Apple has responded in defense, with Natalie Kerris, a spokeswoman for the Cupertino, California-based company saying that:
Meanwhile, Blaze is defending its results. Although it is in fact true that the company was not running a Javascript test, but in fact was running a test against website load times, so in fact a majority of its tests may in fact be valid, the company will not stop being criticized until and unless it can re-run the tests optimally.
Meanwhile, a local test of SunSpider 0.9.1 (in the browser) against an iPad 2 and a Motorola Xoom showed a score of 2056.1 for the iPad 2 and a score of 2062.1 for the Xoom. Lower scores are better, but that is a single pass and most likely means a dead heat.
The admission by Apple means that conspiracy talk can be halted. Since Web apps are a popular way "around" rejections from the App Store (e.g. the official Google Voice app before Apple relented, Readability), some developers theorized that Apple was slowing Web apps on purpose.
This also means that the browser performance tests run by Blaze Software and released on Wednesday were in fact flawed. The site, which offers a free service for measuring mobile web performance, ran a test comparing Android and iOS browsing performance. However it used custom apps for both Android and iOS, and the iOS version used theUIWebView API, which Apple said was lacking in the performance improvements in Safari in iOS 4.3.
The improvements in iOS 4.3 for Safari include the Nitro Javascript engine as well as "certain Safari caches and the browser's 'asynchronous' rendering mode."
Flawed or not, the statements by Apple admitting the differences between the two are likely a response to the report, so that is a good thing. Apple has responded in defense, with Natalie Kerris, a spokeswoman for the Cupertino, California-based company saying that:
“[Apple] regards the tests as flawed because Blaze used its own proprietary application that doesn’t take advantage of Apple Safari browser’s Web-performance optimization. Despite this fundamental testing flaw they still only found an average of a second difference in loading Web pages."Now that Apple has admitted the differences, there may be changes going forward that will equalize the two ways of accessing the Web on iOS.
Meanwhile, Blaze is defending its results. Although it is in fact true that the company was not running a Javascript test, but in fact was running a test against website load times, so in fact a majority of its tests may in fact be valid, the company will not stop being criticized until and unless it can re-run the tests optimally.
Meanwhile, a local test of SunSpider 0.9.1 (in the browser) against an iPad 2 and a Motorola Xoom showed a score of 2056.1 for the iPad 2 and a score of 2062.1 for the Xoom. Lower scores are better, but that is a single pass and most likely means a dead heat.
No comments:
Post a Comment