Narrator in Windows 8 is buggy, sloppily designed and still incomplete, even though it has been greatly improved from the version in Windows 7 which was effectively useless.
Why isn't there a shortcut to move from form element to form element on the Web, like edit boxes, check boxes, etc., similar to all other screen readers? I guess the answer is that with all the unwise choices you made with the shortcut keys, there were no more neighborring letters on the keyboard to use for this functionality. For those that don't know, Narrator's shortcuts are numerous and very hard to remember. It seems that the team made a list of features and then assigned shortcut keys by going through this list and picking neighboring keys on the keyboard to fill it in. Just like that. No logic. No mnemonic combinations were picked.
Its sound effects are not so good compared to Apple's Voiceover, even its TTS is worse than Alex on the Mac. Voiceover has TTSs for over 30 languages all pre-installed on the iPhone/iPad complete with logic to auto-detect and switch TTS when text in another language is encountered. And they don't take up 300 MB like the Windows TTS voices you would have to buy.
Why is apple's Voiceover so much better, with no bugs and fully featured, even since its first iPhone version, whilst Narrator, ten years on, is still so bad. Plus Windows+enter does not work unless you first actually go in and turn this particular shortcut on in Settings. As you can surely understand this is totally useless for a blind person, since if you could go in and turn the shortcut on, you could turn Narrator on as well without the shortcut. The shortcut that actually works is ctrl+Windows+u but it is not documented or perhaps not known even by MS employees.
Voiceover is so so so much more polished and just works in most iPhone/iPad apps. On the other hand, most apps on the Windows Store are inaccessible, such as Skype. I could use Skype on my iPhone since iPhone 3Gs some years ago.
With Microsoft's speed, it would take you ten years to implement what Voiceover v1 had already implemented 4 years ago. Why is that?